1
00:00:04,960 --> 00:00:19,999
[Music]

2
00:00:20,680 --> 00:00:24,800
hi everyone welcome back for the last

3
00:00:22,519 --> 00:00:27,240
talk before we have lunch uh would you

4
00:00:24,800 --> 00:00:29,400
please join me in welcoming to this the

5
00:00:27,240 --> 00:00:31,880
podium uh Clinton Roy who'll be talking

6
00:00:29,400 --> 00:00:34,160
about adding file system contexts to

7
00:00:31,880 --> 00:00:34,160
path

8
00:00:35,559 --> 00:00:40,600
lib it's always helpful to get a

9
00:00:37,800 --> 00:00:43,920
reminder of um what is between you and

10
00:00:40,600 --> 00:00:45,360
lunch so no pressure uh hello everyone

11
00:00:43,920 --> 00:00:47,920
my name is Clinton I'm a software

12
00:00:45,360 --> 00:00:50,520
engineer uh most of my career is has

13
00:00:47,920 --> 00:00:52,399
been around helping researchers uh write

14
00:00:50,520 --> 00:00:55,840
code so that they can do more research

15
00:00:52,399 --> 00:00:57,120
and more science and make life better uh

16
00:00:55,840 --> 00:00:59,359
at the moment I'm at the Australian

17
00:00:57,120 --> 00:01:01,199
synchrotron um a few of my colleagues

18
00:00:59,359 --> 00:01:02,680
have given uh a couple of talks earlier

19
00:01:01,199 --> 00:01:04,360
in the week on the synchron stuff so I

20
00:01:02,680 --> 00:01:06,640
don't have to go over that

21
00:01:04,360 --> 00:01:09,040
fortunately um I'm talking today about

22
00:01:06,640 --> 00:01:11,240
something that is not uh synchrotron

23
00:01:09,040 --> 00:01:13,080
specific uh but it certainly has come up

24
00:01:11,240 --> 00:01:17,159
from uh some of my work at the

25
00:01:13,080 --> 00:01:21,040
synchrotron and it is a very small tweak

26
00:01:17,159 --> 00:01:23,759
to python that I'm suggesting um and you

27
00:01:21,040 --> 00:01:25,479
know as if as with any change the change

28
00:01:23,759 --> 00:01:28,880
is the hardest

29
00:01:25,479 --> 00:01:31,720
part um so during this talk I'm just

30
00:01:28,880 --> 00:01:34,759
going to outline what I'm proposing um

31
00:01:31,720 --> 00:01:36,399
why um how it could be implemented and

32
00:01:34,759 --> 00:01:40,600
what are the next

33
00:01:36,399 --> 00:01:44,280
steps um there is a public part of the

34
00:01:40,600 --> 00:01:48,159
Python Community public website uh uh

35
00:01:44,280 --> 00:01:51,479
called discus or discuss um there is a

36
00:01:48,159 --> 00:01:53,520
section of that called python ideas I'm

37
00:01:51,479 --> 00:01:57,320
trying to choose my words very carefully

38
00:01:53,520 --> 00:02:00,200
here python ideas is a place where ideas

39
00:01:57,320 --> 00:02:01,840
go to get schooled um a lot of people

40
00:02:00,200 --> 00:02:07,560
have ideas they think they're really

41
00:02:01,840 --> 00:02:09,239
good um uh ideas many people have ideas

42
00:02:07,560 --> 00:02:11,319
and they're much like getting a song

43
00:02:09,239 --> 00:02:12,400
stuck in your head they're stuck in your

44
00:02:11,319 --> 00:02:14,800
head they're not stuck in everyone

45
00:02:12,400 --> 00:02:16,280
else's head uh they might sound really

46
00:02:14,800 --> 00:02:19,080
good inside your head but when you sing

47
00:02:16,280 --> 00:02:21,040
them they sound really bad ideas can be

48
00:02:19,080 --> 00:02:23,000
really similar to that they sound really

49
00:02:21,040 --> 00:02:24,879
good in your head but they're not so

50
00:02:23,000 --> 00:02:30,040
crash hot when everyone else looks at

51
00:02:24,879 --> 00:02:31,480
them um there's there's roughly um um I

52
00:02:30,040 --> 00:02:36,519
forget the exact number I think there's

53
00:02:31,480 --> 00:02:40,680
roughly about uh 50 or 60 threads uh of

54
00:02:36,519 --> 00:02:45,159
ideas posted to the python ideas um um

55
00:02:40,680 --> 00:02:47,800
Forum a year and most of those ideas the

56
00:02:45,159 --> 00:02:49,800
response to them is like the last person

57
00:02:47,800 --> 00:02:51,800
who brought it up here's are all the

58
00:02:49,800 --> 00:02:53,519
responses to it that said it's a really

59
00:02:51,800 --> 00:02:57,640
bad idea go and have a look at this

60
00:02:53,519 --> 00:03:00,800
thread or even uh more concretely that

61
00:02:57,640 --> 00:03:03,280
idea um was proposed in the standard

62
00:03:00,800 --> 00:03:06,120
python way uh uh a pep which we'll get

63
00:03:03,280 --> 00:03:07,720
on to later on um and these are all the

64
00:03:06,120 --> 00:03:11,040
reasons that pep didn't

65
00:03:07,720 --> 00:03:12,680
proceed so whenever I have an idea uh

66
00:03:11,040 --> 00:03:16,920
because I subscribe to the python ideas

67
00:03:12,680 --> 00:03:19,040
mailing list um I know how to set my

68
00:03:16,920 --> 00:03:21,959
expectations for any ideas that I come

69
00:03:19,040 --> 00:03:25,920
up with and they are

70
00:03:21,959 --> 00:03:28,000
l so I had an idea and much like a song

71
00:03:25,920 --> 00:03:30,400
I couldn't get it out of my head um and

72
00:03:28,000 --> 00:03:32,760
it was rattling around and matter how

73
00:03:30,400 --> 00:03:33,959
many side projects I said yes to I

74
00:03:32,760 --> 00:03:36,040
couldn't get rid of the idea from my

75
00:03:33,959 --> 00:03:38,000
head so I decided to post it up on

76
00:03:36,040 --> 00:03:39,640
python ideas and let other people shoot

77
00:03:38,000 --> 00:03:41,120
it down so that I could quietly put it

78
00:03:39,640 --> 00:03:43,879
to to

79
00:03:41,120 --> 00:03:45,920
bet um so you know I freely admitted

80
00:03:43,879 --> 00:03:47,920
that I was using a a public resource I

81
00:03:45,920 --> 00:03:49,400
was using other people's time for my own

82
00:03:47,920 --> 00:03:52,079
mental

83
00:03:49,400 --> 00:03:53,480
health um unfortunately someone came

84
00:03:52,079 --> 00:03:55,120
back and said that's kind of an

85
00:03:53,480 --> 00:03:58,159
interesting idea and you should do it

86
00:03:55,120 --> 00:04:01,120
and then he uh uh listed out uh a

87
00:03:58,159 --> 00:04:04,480
technique for doing it

88
00:04:01,120 --> 00:04:07,519
um if you don't know Barney Barney is

89
00:04:04,480 --> 00:04:11,200
the this is the code owners file from uh

90
00:04:07,519 --> 00:04:13,879
cpython GitHub um Barney is the owner of

91
00:04:11,200 --> 00:04:16,720
the path lib Library so unfortunately I

92
00:04:13,879 --> 00:04:18,079
posted an idea and the person who owns

93
00:04:16,720 --> 00:04:19,239
the library said it's an interesting

94
00:04:18,079 --> 00:04:24,400
idea and you should go and have a look

95
00:04:19,239 --> 00:04:29,080
at it so that was that was the first

96
00:04:24,400 --> 00:04:31,759
fail so um I'm not I'm not really

97
00:04:29,080 --> 00:04:34,160
suggesting it is a very small idea as

98
00:04:31,759 --> 00:04:37,720
you you will see uh I'm not suggesting

99
00:04:34,160 --> 00:04:39,360
huge code changes um so most of the

100
00:04:37,720 --> 00:04:43,039
coding examples I'll be showing will be

101
00:04:39,360 --> 00:04:44,840
relatively small um for space reasons

102
00:04:43,039 --> 00:04:47,440
you can assume that all of the examples

103
00:04:44,840 --> 00:04:49,400
have uh this this import here just so

104
00:04:47,440 --> 00:04:51,880
that we can save on Space going forward

105
00:04:49,400 --> 00:04:54,320
so it's a very standard uh import of

106
00:04:51,880 --> 00:04:57,400
path from path

107
00:04:54,320 --> 00:05:00,759
lip um and the change that I'm

108
00:04:57,400 --> 00:05:03,600
suggesting here is there are flags that

109
00:05:00,759 --> 00:05:06,479
are passed to operations in path lib and

110
00:05:03,600 --> 00:05:09,120
the most common ones are to do with case

111
00:05:06,479 --> 00:05:12,240
sensitivity and to following sin links

112
00:05:09,120 --> 00:05:14,919
and I'm suggesting that we move those

113
00:05:12,240 --> 00:05:18,280
flags from the operation so from the

114
00:05:14,919 --> 00:05:23,520
verb and uh we move those flags uh to

115
00:05:18,280 --> 00:05:26,840
the noun uh so to the object so um in

116
00:05:23,520 --> 00:05:29,400
this before case here uh pathlib has a

117
00:05:26,840 --> 00:05:34,199
glob operator so we're asking path lib

118
00:05:29,400 --> 00:05:37,960
to give us all the st. Tiff files and um

119
00:05:34,199 --> 00:05:40,479
if I say case sensitive false um it will

120
00:05:37,960 --> 00:05:43,440
give me all of the star. Tiff lowercase

121
00:05:40,479 --> 00:05:47,400
files and all the star. Tiff uppercase

122
00:05:43,440 --> 00:05:50,600
files um and if one of the elements of

123
00:05:47,400 --> 00:05:53,160
Home Clinton R Pon Au is a Sim link um

124
00:05:50,600 --> 00:05:54,800
I'm telling it to not follow those Sim

125
00:05:53,160 --> 00:05:57,360
links so if there's a Sim Link in that

126
00:05:54,800 --> 00:06:00,319
path don't follow it so all I'm

127
00:05:57,360 --> 00:06:02,919
suggesting is we um

128
00:06:00,319 --> 00:06:06,840
move those flags from the operation to

129
00:06:02,919 --> 00:06:08,639
the object um there is a little bit of a

130
00:06:06,840 --> 00:06:10,400
name difference between the recurse Sim

131
00:06:08,639 --> 00:06:12,800
links and the follow Sim links the

132
00:06:10,400 --> 00:06:14,240
concept is identical there so don't get

133
00:06:12,800 --> 00:06:16,720
hung up on the name of the attribute

134
00:06:14,240 --> 00:06:19,000
changing too much so that that's the

135
00:06:16,720 --> 00:06:22,479
talk so if you need to go off and have

136
00:06:19,000 --> 00:06:25,639
lunch you've got the the guts of the

137
00:06:22,479 --> 00:06:28,360
idea um and I kind of have to describe

138
00:06:25,639 --> 00:06:30,319
why we want to do this now so if your

139
00:06:28,360 --> 00:06:32,120
stomach's nice and comfortable um for

140
00:06:30,319 --> 00:06:35,960
lunch now I'd leave now because this is

141
00:06:32,120 --> 00:06:35,960
where it starts to get a little bit

142
00:06:36,199 --> 00:06:41,190
quasy I am a developer I'm therefore

143
00:06:39,639 --> 00:06:42,240
right

144
00:06:41,190 --> 00:06:45,400
[Music]

145
00:06:42,240 --> 00:06:47,479
so this is this is the essential core

146
00:06:45,400 --> 00:06:49,520
problem when people post their ideas up

147
00:06:47,479 --> 00:06:52,560
to python ideas this is the the core

148
00:06:49,520 --> 00:06:53,840
motivating factor that they've got um

149
00:06:52,560 --> 00:06:57,160
this is the one that you've got to get

150
00:06:53,840 --> 00:07:01,879
over when you post your ideas in

151
00:06:57,160 --> 00:07:03,919
public um I I think long term with path

152
00:07:01,879 --> 00:07:06,759
lib we're just going to start seeing

153
00:07:03,919 --> 00:07:09,160
more and more flags and and I think

154
00:07:06,759 --> 00:07:11,360
those flags are good uh so so having

155
00:07:09,160 --> 00:07:13,280
having those flags uh to be able to

156
00:07:11,360 --> 00:07:15,400
change the way that operations happen I

157
00:07:13,280 --> 00:07:17,800
think is a good thing but I think it

158
00:07:15,400 --> 00:07:20,039
might be a better thing if we have one

159
00:07:17,800 --> 00:07:21,759
place where we put all of those flags

160
00:07:20,039 --> 00:07:22,560
rather than adding Flags to lots of

161
00:07:21,759 --> 00:07:24,240
different

162
00:07:22,560 --> 00:07:26,840
operations and that's that's a little

163
00:07:24,240 --> 00:07:30,240
bit of a taste thing uh but I also think

164
00:07:26,840 --> 00:07:32,720
it means that when we add a flag we are

165
00:07:30,240 --> 00:07:34,720
not uh just adding a flag to one

166
00:07:32,720 --> 00:07:38,720
operation we're thinking about how that

167
00:07:34,720 --> 00:07:41,160
new flag can uh alter the behavior of

168
00:07:38,720 --> 00:07:43,520
the entire path lip uh

169
00:07:41,160 --> 00:07:46,840
Library um and it means that if we add a

170
00:07:43,520 --> 00:07:49,919
new flag we're adding it in one place so

171
00:07:46,840 --> 00:07:53,000
the documentation for the path lib init

172
00:07:49,919 --> 00:07:55,759
uh function would grow all of the uh

173
00:07:53,000 --> 00:08:00,120
Flags but none of the other operations

174
00:07:55,759 --> 00:08:00,120
would grow uh in their complexity

175
00:08:00,680 --> 00:08:08,919
um the the main thing I think here is

176
00:08:03,599 --> 00:08:12,080
that it's very possible to have paths as

177
00:08:08,919 --> 00:08:16,919
um parts of an API where you're passing

178
00:08:12,080 --> 00:08:18,319
a directory along to an API and that API

179
00:08:16,919 --> 00:08:20,520
you don't want to have to give it all of

180
00:08:18,319 --> 00:08:23,599
these extra flags as part of that API

181
00:08:20,520 --> 00:08:25,919
because the path is really just a input

182
00:08:23,599 --> 00:08:28,840
or output operator and you don't want

183
00:08:25,919 --> 00:08:31,240
your non-path API to have to deal with

184
00:08:28,840 --> 00:08:33,519
those flags

185
00:08:31,240 --> 00:08:36,560
so this is where the work side of things

186
00:08:33,519 --> 00:08:38,440
come in let's say I have a science API

187
00:08:36,560 --> 00:08:40,880
that I give a directory and that

188
00:08:38,440 --> 00:08:44,039
directory might have oh I don't know

189
00:08:40,880 --> 00:08:44,959
hundreds of Tiff files from a camera

190
00:08:44,039 --> 00:08:49,399
that I

191
00:08:44,959 --> 00:08:51,440
control um at the moment my science API

192
00:08:49,399 --> 00:08:54,360
would if if there were sings in that

193
00:08:51,440 --> 00:08:55,720
path because it was a nested NFS Mount

194
00:08:54,360 --> 00:08:58,000
from three servers away and there were

195
00:08:55,720 --> 00:09:00,760
lots of Zim links in that chain at the

196
00:08:58,000 --> 00:09:02,720
moment I would have to tell my science

197
00:09:00,760 --> 00:09:05,160
API that doesn't really care about Sim

198
00:09:02,720 --> 00:09:10,360
links that it needs to follow the Sim

199
00:09:05,160 --> 00:09:13,040
links so if I get my change in uh that

200
00:09:10,360 --> 00:09:15,920
op option would get passed along to the

201
00:09:13,040 --> 00:09:17,360
path object the science API wouldn't

202
00:09:15,920 --> 00:09:19,279
have to think about Sim links wouldn't

203
00:09:17,360 --> 00:09:22,360
have to know about Sim links uh but it

204
00:09:19,279 --> 00:09:24,959
would get modified uh by those flags and

205
00:09:22,360 --> 00:09:28,000
in my mind that's appropriate because my

206
00:09:24,959 --> 00:09:29,079
science API doesn't know what SIM links

207
00:09:28,000 --> 00:09:29,920
are shouldn't have to know what SIM

208
00:09:29,079 --> 00:09:31,560
links are

209
00:09:29,920 --> 00:09:33,680
so it's that separation of concerns in

210
00:09:31,560 --> 00:09:33,680
the

211
00:09:33,920 --> 00:09:41,399
API um so as long as well as uh Barney

212
00:09:39,160 --> 00:09:43,519
saying it's it's an interesting idea he

213
00:09:41,399 --> 00:09:46,399
also gave me a complete recipe for had

214
00:09:43,519 --> 00:09:47,600
to go about implementing a prototype so

215
00:09:46,399 --> 00:09:49,120
the rest of the talk is going through

216
00:09:47,600 --> 00:09:52,760
that

217
00:09:49,120 --> 00:09:54,519
prototype um so I am making a class

218
00:09:52,760 --> 00:09:58,560
called context path because all of these

219
00:09:54,519 --> 00:10:01,079
flags are context in a certain way um in

220
00:09:58,560 --> 00:10:02,920
the fullness of time if this idea was

221
00:10:01,079 --> 00:10:05,519
accepted there wouldn't be a context

222
00:10:02,920 --> 00:10:07,839
path it would go into path itself but

223
00:10:05,519 --> 00:10:10,760
for this test run we're making a new

224
00:10:07,839 --> 00:10:13,680
class called context path the really

225
00:10:10,760 --> 00:10:18,200
important thing here is that in Python

226
00:10:13,680 --> 00:10:22,160
312 and 313 Barnie's done a lot of work

227
00:10:18,200 --> 00:10:24,640
to rejig path lib so that you can subass

228
00:10:22,160 --> 00:10:26,480
from path and do so in a really useful

229
00:10:24,640 --> 00:10:28,800
manner so if you want to do experiments

230
00:10:26,480 --> 00:10:32,160
around path and path lib it is really

231
00:10:28,800 --> 00:10:33,640
easy to do do um so much so that he's

232
00:10:32,160 --> 00:10:35,040
kind of got a rot response about how to

233
00:10:33,640 --> 00:10:37,480
try these things

234
00:10:35,040 --> 00:10:41,880
out

235
00:10:37,480 --> 00:10:44,680
so um the second implementation step is

236
00:10:41,880 --> 00:10:48,760
passing along my my context Flags to the

237
00:10:44,680 --> 00:10:52,880
init uh operation so my init operation

238
00:10:48,760 --> 00:10:56,079
here uh uh very standard python where

239
00:10:52,880 --> 00:10:59,440
I've got uh my star ARS which is all of

240
00:10:56,079 --> 00:11:02,120
the parts of the path

241
00:10:59,440 --> 00:11:06,360
um I've got my case sensitivity flag my

242
00:11:02,120 --> 00:11:08,880
follow Sim links flag I call uh super of

243
00:11:06,360 --> 00:11:11,320
in it so that's the the path in it so

244
00:11:08,880 --> 00:11:14,120
that initializes path how it always has

245
00:11:11,320 --> 00:11:18,760
been and then I store my Flags in the

246
00:11:14,120 --> 00:11:21,480
object so very um very normal standard

247
00:11:18,760 --> 00:11:25,200
basic python

248
00:11:21,480 --> 00:11:28,160
there uh step three uh is Implement with

249
00:11:25,200 --> 00:11:32,760
segments with segments is the Workhorse

250
00:11:28,160 --> 00:11:35,480
of the path uh path lid path object uh

251
00:11:32,760 --> 00:11:38,600
anytime you are making A New Path based

252
00:11:35,480 --> 00:11:39,959
on uh a previous path like you're using

253
00:11:38,600 --> 00:11:41,639
the slash operator to add a new

254
00:11:39,959 --> 00:11:44,360
directory or file name into it it

255
00:11:41,639 --> 00:11:49,360
eventually gets called down to with

256
00:11:44,360 --> 00:11:50,839
segments so the um the path

257
00:11:49,360 --> 00:11:55,480
implementation of withd segments is

258
00:11:50,839 --> 00:11:58,800
really quite clever it grabs self asks

259
00:11:55,480 --> 00:12:01,200
me what the type of self is and in path

260
00:11:58,800 --> 00:12:04,040
that answer is path and then it passes

261
00:12:01,200 --> 00:12:06,480
along those path segments to its own

262
00:12:04,040 --> 00:12:10,160
class so it's calling in it again

263
00:12:06,480 --> 00:12:12,680
essentially so my context uh path

264
00:12:10,160 --> 00:12:14,279
implementation very very similar does

265
00:12:12,680 --> 00:12:17,240
exactly the same thing but just passes

266
00:12:14,279 --> 00:12:19,560
those flags along so I'm doing exactly

267
00:12:17,240 --> 00:12:21,440
the same thing as the original path

268
00:12:19,560 --> 00:12:24,079
class but I'm just passing my context

269
00:12:21,440 --> 00:12:24,079
Flags

270
00:12:25,399 --> 00:12:28,959
along I can't hear anyone's stomach

271
00:12:27,680 --> 00:12:30,560
rumbling or anything like that so that's

272
00:12:28,959 --> 00:12:32,240
good

273
00:12:30,560 --> 00:12:34,000
um so that's essentially the

274
00:12:32,240 --> 00:12:39,199
implementation uh like a test

275
00:12:34,000 --> 00:12:42,600
implementation of my idea um in terms of

276
00:12:39,199 --> 00:12:45,079
uh creating a context path passing the

277
00:12:42,600 --> 00:12:48,160
flags along now it's time to use the

278
00:12:45,079 --> 00:12:49,360
flags I'm not um like like the really

279
00:12:48,160 --> 00:12:51,560
important thing here is I'm not talking

280
00:12:49,360 --> 00:12:53,480
about changing the behavior of path lib

281
00:12:51,560 --> 00:12:58,240
I'm just talking about changing the way

282
00:12:53,480 --> 00:13:00,240
you can figure uh path lib so um one of

283
00:12:58,240 --> 00:13:03,720
the operations that that path lib does

284
00:13:00,240 --> 00:13:05,480
is a stat uh a stat uh asks for

285
00:13:03,720 --> 00:13:06,800
information about a file or directory

286
00:13:05,480 --> 00:13:12,720
that you're you're pointing

287
00:13:06,800 --> 00:13:15,040
to so um the implementation um of stat I

288
00:13:12,720 --> 00:13:18,040
can call that in my context uh

289
00:13:15,040 --> 00:13:22,560
implementation of stat by calling super.

290
00:13:18,040 --> 00:13:25,920
stat and when I pass so stat already has

291
00:13:22,560 --> 00:13:28,120
a uh a flag on the verb to follow Sim

292
00:13:25,920 --> 00:13:32,600
links or not so all I do in my

293
00:13:28,120 --> 00:13:35,360
implementation is grab my uh uh

294
00:13:32,600 --> 00:13:37,880
internally stored flag self. follow Sim

295
00:13:35,360 --> 00:13:40,560
links so I'm not changing the operations

296
00:13:37,880 --> 00:13:43,480
of anything it's just um how and where

297
00:13:40,560 --> 00:13:46,480
the flags are stored and passed

298
00:13:43,480 --> 00:13:48,440
along uh and similarly with glob where

299
00:13:46,480 --> 00:13:52,600
we're looking at globbing start. Tiff

300
00:13:48,440 --> 00:13:55,279
files before very similar um I call the

301
00:13:52,600 --> 00:13:57,199
current implementation of glob but I

302
00:13:55,279 --> 00:13:59,199
just pass along the case sensitive and

303
00:13:57,199 --> 00:14:04,480
follow Sim link flags for my own

304
00:13:59,199 --> 00:14:09,600
internal uh flag uh registry

305
00:14:04,480 --> 00:14:13,440
memory so to think about um how this

306
00:14:09,600 --> 00:14:16,600
would operate for testing we have a

307
00:14:13,440 --> 00:14:19,360
testing directory here um we've got two

308
00:14:16,600 --> 00:14:22,360
Tiff files here one's lowcase one's

309
00:14:19,360 --> 00:14:23,920
uppercase so when we switch the case

310
00:14:22,360 --> 00:14:26,680
sensitive flag and we do a glob

311
00:14:23,920 --> 00:14:31,279
operation we'll expect to either get one

312
00:14:26,680 --> 00:14:33,639
or two Tiff files there and we' got uh a

313
00:14:31,279 --> 00:14:37,560
destination D at the top and we've also

314
00:14:33,639 --> 00:14:39,440
got a Sim link um uh pointing to that D

315
00:14:37,560 --> 00:14:44,440
at the bottom

316
00:14:39,440 --> 00:14:46,480
um don't worry the context um important

317
00:14:44,440 --> 00:14:48,759
parts of this slide will be presented

318
00:14:46,480 --> 00:14:50,360
further on but that's I've only got four

319
00:14:48,759 --> 00:14:53,160
files that I'm testing with just as a

320
00:14:50,360 --> 00:14:53,160
proof of concept

321
00:14:53,639 --> 00:15:01,040
test so now we look at some basic tests

322
00:14:57,639 --> 00:15:04,480
um for context path so again these are

323
00:15:01,040 --> 00:15:06,480
common Imports so stat is a standalone

324
00:15:04,480 --> 00:15:09,480
uh python module uh it's a built-in

325
00:15:06,480 --> 00:15:11,279
python module when you do a stat uh on a

326
00:15:09,480 --> 00:15:14,079
file or directory you get back a whole

327
00:15:11,279 --> 00:15:15,800
bunch of numbers stat essentially has a

328
00:15:14,079 --> 00:15:20,959
whole bunch of constants that let you

329
00:15:15,800 --> 00:15:22,639
ask um is this file thing a file is it a

330
00:15:20,959 --> 00:15:25,639
Sim link is a

331
00:15:22,639 --> 00:15:27,240
directory and for my tests I import

332
00:15:25,639 --> 00:15:30,759
context path as

333
00:15:27,240 --> 00:15:32,639
well uh so these are just Imports that

334
00:15:30,759 --> 00:15:34,399
are repeated for all the tests and I've

335
00:15:32,639 --> 00:15:36,199
just put them up here to save space

336
00:15:34,399 --> 00:15:41,040
further

337
00:15:36,199 --> 00:15:44,759
on so this is a case sensitive test so

338
00:15:41,040 --> 00:15:46,839
this is the subset of the test area that

339
00:15:44,759 --> 00:15:48,399
um we're interested in here so we got

340
00:15:46,839 --> 00:15:49,399
the lowercase Tiff file the uppercase

341
00:15:48,399 --> 00:15:53,240
Tiff

342
00:15:49,399 --> 00:15:55,800
file um we create our context path we

343
00:15:53,240 --> 00:15:59,759
pass along case sensitive true so we

344
00:15:55,800 --> 00:16:02,560
care about the case when we do um a glob

345
00:15:59,759 --> 00:16:05,440
on star. Tiff and that is a lower

346
00:16:02,560 --> 00:16:08,399
lowercase do Tiff we expect that we'll

347
00:16:05,440 --> 00:16:12,600
only get the lowercase image. Tiff file

348
00:16:08,399 --> 00:16:16,639
back so that's our case sensitive uh

349
00:16:12,600 --> 00:16:20,440
test there and very

350
00:16:16,639 --> 00:16:24,120
similarly same um area with the two Tiff

351
00:16:20,440 --> 00:16:27,319
files uppercase and lower case this time

352
00:16:24,120 --> 00:16:30,079
we do we set the case sensitive option

353
00:16:27,319 --> 00:16:33,000
to false so we don't care about case so

354
00:16:30,079 --> 00:16:35,720
this time when we do a glob on star.

355
00:16:33,000 --> 00:16:38,079
Tiff it will glob on both the lower case

356
00:16:35,720 --> 00:16:40,440
and the uppercase you can see the assert

357
00:16:38,079 --> 00:16:43,120
there we expect to get those two files

358
00:16:40,440 --> 00:16:45,720
back so that's the the test so we've

359
00:16:43,120 --> 00:16:47,839
tested uh the on and the off parts of

360
00:16:45,720 --> 00:16:47,839
the

361
00:16:48,160 --> 00:16:55,639
flag um Sim links are a little bit um

362
00:16:53,800 --> 00:16:59,040
little bit harder to reason about to

363
00:16:55,639 --> 00:17:00,720
think about um I'm a Unix person and I

364
00:16:59,040 --> 00:17:04,280
know through many years of experience

365
00:17:00,720 --> 00:17:09,039
that I get them wrong all the time um so

366
00:17:04,280 --> 00:17:10,559
this is the um this is the the test for

367
00:17:09,039 --> 00:17:13,079
uh following Sim links when we're doing

368
00:17:10,559 --> 00:17:14,919
a stat operation the subset of the test

369
00:17:13,079 --> 00:17:17,319
area that we're interested in here is

370
00:17:14,919 --> 00:17:19,400
the destination dur so it's a directory

371
00:17:17,319 --> 00:17:23,199
and the Sim link that points to that

372
00:17:19,400 --> 00:17:26,880
directory we make a context path to our

373
00:17:23,199 --> 00:17:29,360
Sim link and we say uh follow the Sim

374
00:17:26,880 --> 00:17:32,000
links um and then we do the St

375
00:17:29,360 --> 00:17:34,600
on that so because we're following the

376
00:17:32,000 --> 00:17:36,240
Sim links when we do the stat we won't

377
00:17:34,600 --> 00:17:39,200
be doing the stat on the Sim link we'll

378
00:17:36,240 --> 00:17:43,440
be doing it on the destination directory

379
00:17:39,200 --> 00:17:45,760
so when we ask the stat results is it a

380
00:17:43,440 --> 00:17:47,039
Sim link it'll be false because we're

381
00:17:45,760 --> 00:17:49,880
following the Sim links so we'll

382
00:17:47,039 --> 00:17:52,520
actually hit the directory and when we

383
00:17:49,880 --> 00:17:54,280
ask the the state results is it a

384
00:17:52,520 --> 00:17:55,600
directory because we followed the Sim

385
00:17:54,280 --> 00:17:56,720
links and we've hit the directory it

386
00:17:55,600 --> 00:18:00,400
will

387
00:17:56,720 --> 00:18:01,880
be and similarly where we're in a mode

388
00:18:00,400 --> 00:18:04,159
where we don't want to follow the Sim

389
00:18:01,880 --> 00:18:07,120
links where we want to Halt on the Sim

390
00:18:04,159 --> 00:18:10,120
links um it's kind of

391
00:18:07,120 --> 00:18:13,400
Switched so we pass along follow Sim

392
00:18:10,120 --> 00:18:16,919
links to be false um which means that

393
00:18:13,400 --> 00:18:18,880
when we're doing the stat um when we're

394
00:18:16,919 --> 00:18:21,840
uh realizing that path it'll stop at the

395
00:18:18,880 --> 00:18:24,960
Sim link so when we check the state

396
00:18:21,840 --> 00:18:26,919
result we can see that it is a Sim link

397
00:18:24,960 --> 00:18:30,120
and that it is not a d because it hasn't

398
00:18:26,919 --> 00:18:32,840
followed that Sim link through um

399
00:18:30,120 --> 00:18:34,320
and I don't think I'll tell you how many

400
00:18:32,840 --> 00:18:36,520
times I had to do this test until I got

401
00:18:34,320 --> 00:18:36,520
it

402
00:18:36,559 --> 00:18:42,440
right um so I've got the idea I've got a

403
00:18:40,600 --> 00:18:45,520
very basic implementation I've got a

404
00:18:42,440 --> 00:18:46,880
conference talk out of it as well um so

405
00:18:45,520 --> 00:18:48,480
what's the next thing to do now that

406
00:18:46,880 --> 00:18:51,039
I've got a really basic implementation

407
00:18:48,480 --> 00:18:54,200
of the idea I need to get back to Barney

408
00:18:51,039 --> 00:18:56,280
um show him the code see what he thinks

409
00:18:54,200 --> 00:18:58,200
um and and that's kind of like the

410
00:18:56,280 --> 00:19:02,600
code's the really easy part like the the

411
00:18:58,200 --> 00:19:05,480
point of this talk is to to show that um

412
00:19:02,600 --> 00:19:07,000
you know software a lot of the time is

413
00:19:05,480 --> 00:19:08,559
really easy and it's dealing with the

414
00:19:07,000 --> 00:19:13,240
people and the change management that's

415
00:19:08,559 --> 00:19:15,720
really hard so the next uh so assuming

416
00:19:13,240 --> 00:19:17,799
that that Barney likes the code and

417
00:19:15,720 --> 00:19:19,360
wants to go ahead now we have to think

418
00:19:17,799 --> 00:19:21,960
about how we're going to incorporate

419
00:19:19,360 --> 00:19:25,200
this change into the py python

420
00:19:21,960 --> 00:19:28,360
ecosystem um the structured way of doing

421
00:19:25,200 --> 00:19:32,000
this uh in the python Community is uh

422
00:19:28,360 --> 00:19:35,760
with a python enhancement uh proposal or

423
00:19:32,000 --> 00:19:37,360
pep um it's a structured document um

424
00:19:35,760 --> 00:19:41,120
which talks about all of the technical

425
00:19:37,360 --> 00:19:43,880
challenges of the idea uh it explicitly

426
00:19:41,120 --> 00:19:46,679
brings out um how we're going to teach

427
00:19:43,880 --> 00:19:48,960
this change because uh at the roots of

428
00:19:46,679 --> 00:19:51,200
python uh it was designed to be a

429
00:19:48,960 --> 00:19:52,720
language that's easy to teach so if

430
00:19:51,200 --> 00:19:55,400
you're making a change to the language

431
00:19:52,720 --> 00:19:57,960
you have to um outline how you're going

432
00:19:55,400 --> 00:20:02,640
to teach this change uh which I think is

433
00:19:57,960 --> 00:20:03,760
uh very useful thing um but there's also

434
00:20:02,640 --> 00:20:06,360
uh thinking about backwards

435
00:20:03,760 --> 00:20:09,480
compatibility because it's a change so

436
00:20:06,360 --> 00:20:11,159
it's it's we're going to um you know if

437
00:20:09,480 --> 00:20:12,919
this gets up it's going to change how

438
00:20:11,159 --> 00:20:15,039
people use path lib and we have to think

439
00:20:12,919 --> 00:20:18,280
about all of the code that's out there

440
00:20:15,039 --> 00:20:19,840
that uses these flags on the verb and

441
00:20:18,280 --> 00:20:22,440
how are we're going to migrate to having

442
00:20:19,840 --> 00:20:25,280
the flags on the noun do we have the

443
00:20:22,440 --> 00:20:27,039
flags in both places do we deprecate

444
00:20:25,280 --> 00:20:30,080
them do we think about eventually

445
00:20:27,039 --> 00:20:32,320
removing them do we use the Tri-State

446
00:20:30,080 --> 00:20:35,120
nature of these flags where by default

447
00:20:32,320 --> 00:20:37,400
they're none um you can actively set

448
00:20:35,120 --> 00:20:39,960
them to true or false if one of the

449
00:20:37,400 --> 00:20:42,640
flags is none and one of the other flags

450
00:20:39,960 --> 00:20:45,000
is uh true or false does that override

451
00:20:42,640 --> 00:20:46,360
things or does the verb always override

452
00:20:45,000 --> 00:20:48,600
the noun or does the noun always

453
00:20:46,360 --> 00:20:51,520
override the verb there's lots of

454
00:20:48,600 --> 00:20:53,520
options and fortunately um I won't be in

455
00:20:51,520 --> 00:20:59,760
charge of making that

456
00:20:53,520 --> 00:21:03,400
decision um so if we go down this track

457
00:20:59,760 --> 00:21:08,000
um it's it's it's one of these changes

458
00:21:03,400 --> 00:21:09,600
where it sort of uh opens the door just

459
00:21:08,000 --> 00:21:11,720
a crack and it can sort of see through

460
00:21:09,600 --> 00:21:15,880
all the things that you would do with it

461
00:21:11,720 --> 00:21:17,600
so um the buzzing in the brain this is

462
00:21:15,880 --> 00:21:21,640
like the flapping in the brain versus

463
00:21:17,600 --> 00:21:24,600
the buzzing in the brain so Windows has

464
00:21:21,640 --> 00:21:26,480
things that behave like Sim links but

465
00:21:24,600 --> 00:21:29,159
regular users don't have the Privileges

466
00:21:26,480 --> 00:21:31,480
to use them they do have things called

467
00:21:29,159 --> 00:21:34,919
Junctions which are kind of like Sim

468
00:21:31,480 --> 00:21:36,279
links but they're only um if I'm getting

469
00:21:34,919 --> 00:21:38,880
if I'm remembering

470
00:21:36,279 --> 00:21:40,840
correctly uh Junctions are Sim links

471
00:21:38,880 --> 00:21:42,080
that you can only point to directories

472
00:21:40,840 --> 00:21:44,600
and you can only Point directories on

473
00:21:42,080 --> 00:21:47,799
the same file system there's already

474
00:21:44,600 --> 00:21:51,200
code inside of path lib to check if

475
00:21:47,799 --> 00:21:54,559
something's a junction and part of part

476
00:21:51,200 --> 00:21:58,600
of my um impetus for doing this is

477
00:21:54,559 --> 00:22:00,559
instead of adding a follow Junction flag

478
00:21:58,600 --> 00:22:03,039
to all of the calls that it makes sense

479
00:22:00,559 --> 00:22:06,640
we just add that follow Junction flag

480
00:22:03,039 --> 00:22:10,760
just in the init function and we call it

481
00:22:06,640 --> 00:22:11,960
done um there's lots of fun things with

482
00:22:10,760 --> 00:22:16,360
lots of different Windows operating

483
00:22:11,960 --> 00:22:19,600
systems where the path can be encoded uh

484
00:22:16,360 --> 00:22:21,840
with different um Unicode mappings and

485
00:22:19,600 --> 00:22:25,039
different you know pre- Unicode mappings

486
00:22:21,840 --> 00:22:27,320
so that you can have uh nicely spelled

487
00:22:25,039 --> 00:22:30,360
uh human names in file names and

488
00:22:27,320 --> 00:22:32,159
directories um at the moment python sort

489
00:22:30,360 --> 00:22:34,559
of takes a punt on that and says we

490
00:22:32,159 --> 00:22:37,720
always do this which is totally fine

491
00:22:34,559 --> 00:22:39,960
totally legit way of doing things but if

492
00:22:37,720 --> 00:22:42,320
we had options on path we could set

493
00:22:39,960 --> 00:22:45,360
those options and we could get nicely

494
00:22:42,320 --> 00:22:47,200
printed names out of paths on Windows

495
00:22:45,360 --> 00:22:48,720
that our users like to read without

496
00:22:47,200 --> 00:22:52,200
having to do any

497
00:22:48,720 --> 00:22:55,440
work um one of the things with Sim links

498
00:22:52,200 --> 00:22:58,640
is that it's very easy to do a really

499
00:22:55,440 --> 00:22:59,760
long chain of sim links ping-ponging all

500
00:22:58,640 --> 00:23:02,480
around your file

501
00:22:59,760 --> 00:23:05,919
system uh you can get into Loops quite

502
00:23:02,480 --> 00:23:07,440
easily if you uh if you forget Which Way

503
00:23:05,919 --> 00:23:09,840
Around The Operators to creating Sim

504
00:23:07,440 --> 00:23:14,720
link uh go which you know I never did

505
00:23:09,840 --> 00:23:16,880
that anymore um so it's very possible to

506
00:23:14,720 --> 00:23:20,760
request following Sim links attempt to

507
00:23:16,880 --> 00:23:22,760
open up a file on a Sim link and just go

508
00:23:20,760 --> 00:23:26,159
round and round in a loop now if we let

509
00:23:22,760 --> 00:23:28,520
that happen um that would be a nice way

510
00:23:26,159 --> 00:23:32,240
of doing a denial of service attack on

511
00:23:28,520 --> 00:23:35,559
on some code so the security people have

512
00:23:32,240 --> 00:23:38,440
have added limits to uh the underlying

513
00:23:35,559 --> 00:23:40,600
file system code that says um if you're

514
00:23:38,440 --> 00:23:42,960
following Sim links and you hit 100 Sim

515
00:23:40,600 --> 00:23:45,679
links that's it we fail and you can't

516
00:23:42,960 --> 00:23:47,279
open that file wouldn't it be nice if at

517
00:23:45,679 --> 00:23:49,840
the python layer we could do something

518
00:23:47,279 --> 00:23:51,919
similar and say you can only follow 10

519
00:23:49,840 --> 00:23:54,400
Sim Links at the at the at the python

520
00:23:51,919 --> 00:23:56,679
layer so that would be an interesting

521
00:23:54,400 --> 00:23:58,919
way of doing security stuff at the path

522
00:23:56,679 --> 00:24:01,679
liway

523
00:23:58,919 --> 00:24:03,320
um there's lots of things to think about

524
00:24:01,679 --> 00:24:05,080
from the backwards compatibility point

525
00:24:03,320 --> 00:24:07,279
of view and and I've already mentioned

526
00:24:05,080 --> 00:24:09,279
this like what would take prec what

527
00:24:07,279 --> 00:24:13,400
would take precedence would the the noun

528
00:24:09,279 --> 00:24:15,159
or the verb Flags take precedence um uh

529
00:24:13,400 --> 00:24:19,320
do we use the Tri-State nature of these

530
00:24:15,159 --> 00:24:22,000
flags to uh make things nice um would we

531
00:24:19,320 --> 00:24:24,400
actually deprecate the arguments uh on

532
00:24:22,000 --> 00:24:26,440
the verbs or would we just leave them

533
00:24:24,400 --> 00:24:27,559
there as an optional extra for those

534
00:24:26,440 --> 00:24:29,240
times when you do actually want to

535
00:24:27,559 --> 00:24:31,919
overwrite

536
00:24:29,240 --> 00:24:35,080
um if you're writing a program or an

537
00:24:31,919 --> 00:24:38,000
application and you've got lots of paths

538
00:24:35,080 --> 00:24:40,640
coming in and out maybe for that

539
00:24:38,000 --> 00:24:43,520
application or Library you want to

540
00:24:40,640 --> 00:24:46,000
always have the same flags for all of

541
00:24:43,520 --> 00:24:49,200
the paths that you create so maybe we

542
00:24:46,000 --> 00:24:50,880
provide a path Factory that sets all of

543
00:24:49,200 --> 00:24:53,039
those flags maybe that's something to

544
00:24:50,880 --> 00:24:53,039
think

545
00:24:55,520 --> 00:25:05,559
about um so that's that's ESS the talk

546
00:24:59,720 --> 00:25:10,600
um I like using old technology um so uh

547
00:25:05,559 --> 00:25:15,080
this slid uh deck um was made with emac

548
00:25:10,600 --> 00:25:16,960
and org mode and uh latch and TI Zed um

549
00:25:15,080 --> 00:25:19,600
the new thing for me is I've actually

550
00:25:16,960 --> 00:25:21,520
paid for a type face which is a sentence

551
00:25:19,600 --> 00:25:23,600
that I never thought I would say but the

552
00:25:21,520 --> 00:25:27,240
uh the monotype face there is uh

553
00:25:23,600 --> 00:25:29,120
courtesy of berley mono and with that I

554
00:25:27,240 --> 00:25:34,320
have time for questions

555
00:25:29,120 --> 00:25:37,360
[Applause]

556
00:25:34,320 --> 00:25:40,360
and who's up for the first question

557
00:25:37,360 --> 00:25:40,360
yeah

558
00:25:41,120 --> 00:25:48,399
no come on you knew i' have something um

559
00:25:45,799 --> 00:25:50,240
so you you already answered my first

560
00:25:48,399 --> 00:25:52,200
question which was the default flag

561
00:25:50,240 --> 00:25:55,600
model um but the other thing that was

562
00:25:52,200 --> 00:25:57,880
giving me Vibes of was decimal contexts

563
00:25:55,600 --> 00:25:58,960
and whether you'd considered the decimal

564
00:25:57,880 --> 00:26:00,960
context

565
00:25:58,960 --> 00:26:04,440
type

566
00:26:00,960 --> 00:26:09,080
approach yeah so

567
00:26:04,440 --> 00:26:10,960
um so similarly so programming is you

568
00:26:09,080 --> 00:26:12,120
know not all the time but a lot of the

569
00:26:10,960 --> 00:26:15,279
time it's

570
00:26:12,120 --> 00:26:17,720
about uh making assumptions versus

571
00:26:15,279 --> 00:26:20,520
knowing the assumptions that you've made

572
00:26:17,720 --> 00:26:23,120
so we've done that in a lot of

573
00:26:20,520 --> 00:26:25,760
contexts and one of these context is

574
00:26:23,120 --> 00:26:27,720
with decimal numbers uh so there's a

575
00:26:25,760 --> 00:26:29,760
whole bunch of flags that you can set

576
00:26:27,720 --> 00:26:32,200
around dealing with python decimal

577
00:26:29,760 --> 00:26:34,720
numbers and this is not a python

578
00:26:32,200 --> 00:26:36,600
specific thing this is the the standard

579
00:26:34,720 --> 00:26:40,399
uh for decimal numbers that the python

580
00:26:36,600 --> 00:26:43,000
mation bases itself on but in order to

581
00:26:40,399 --> 00:26:46,480
make life easy we hid a lot of those

582
00:26:43,000 --> 00:26:48,679
details inside the module um now I might

583
00:26:46,480 --> 00:26:55,720
be getting this wrong so so shout me

584
00:26:48,679 --> 00:26:59,039
down I am um it's it's difficult because

585
00:26:55,720 --> 00:27:01,039
those contexts are built into the module

586
00:26:59,039 --> 00:27:02,480
it's a little bit ugly to change those

587
00:27:01,039 --> 00:27:05,679
contexts for two different decimal

588
00:27:02,480 --> 00:27:09,799
numbers and get them to talk together

589
00:27:05,679 --> 00:27:11,880
nicely so I like I I don't know like in

590
00:27:09,799 --> 00:27:16,960
my mind the the decimal context way of

591
00:27:11,880 --> 00:27:18,520
doing things is not great and and maybe

592
00:27:16,960 --> 00:27:22,640
shouldn't be a recipe for how we do this

593
00:27:18,520 --> 00:27:24,760
in the future but you know that's uh in

594
00:27:22,640 --> 00:27:27,399
an attempt to make decimal numbers as

595
00:27:24,760 --> 00:27:29,520
easy and as nice to deal with uh uh

596
00:27:27,399 --> 00:27:31,600
integer literals we hid some of the

597
00:27:29,520 --> 00:27:33,559
context it's still available and you can

598
00:27:31,600 --> 00:27:36,760
still get it but it's not as nice as it

599
00:27:33,559 --> 00:27:38,440
could be it I I I don't know if you

600
00:27:36,760 --> 00:27:43,799
agree with that or not

601
00:27:38,440 --> 00:27:46,360
but yes yes and and fortunately I don't

602
00:27:43,799 --> 00:27:50,159
have to make that decision think we got

603
00:27:46,360 --> 00:27:50,159
one right at the back oh

604
00:27:52,720 --> 00:27:57,080
yep I'm glad that the video operator

605
00:27:55,200 --> 00:27:58,480
didn't invent a cute way of

606
00:27:57,080 --> 00:28:00,399
communicating to me with the video

607
00:27:58,480 --> 00:28:02,360
recorder I don't have a mic they didn't

608
00:28:00,399 --> 00:28:04,240
give me one they don't trust me with it

609
00:28:02,360 --> 00:28:06,519
uh I wrote my my question down because I

610
00:28:04,240 --> 00:28:10,120
forgot it immediately uh I know you told

611
00:28:06,519 --> 00:28:13,039
us to disregard the uh rename of rer sim

612
00:28:10,120 --> 00:28:17,320
links to follow links yes is that rename

613
00:28:13,039 --> 00:28:19,120
partially so that it is easier to um

614
00:28:17,320 --> 00:28:21,200
introduce that number limit or is it

615
00:28:19,120 --> 00:28:23,600
also just to make it clearer that it is

616
00:28:21,200 --> 00:28:27,600
stopping any any Sim links rather than

617
00:28:23,600 --> 00:28:30,320
like yeah yeah yeah so it's it's very

618
00:28:27,600 --> 00:28:33,960
possible that we would need to think

619
00:28:30,320 --> 00:28:36,519
well we would need to think about the uh

620
00:28:33,960 --> 00:28:38,159
naming of these flags if we change the

621
00:28:36,519 --> 00:28:40,880
location of the flags we need to think

622
00:28:38,159 --> 00:28:44,880
very carefully about the naming of them

623
00:28:40,880 --> 00:28:47,960
uh so it is entirely possible that

624
00:28:44,880 --> 00:28:51,799
without me thinking of a explicit case

625
00:28:47,960 --> 00:28:53,480
of it that we might have two sink flags

626
00:28:51,799 --> 00:28:55,240
that mean slightly different things for

627
00:28:53,480 --> 00:28:59,000
different operators that's entirely

628
00:28:55,240 --> 00:29:00,960
possible um and that option is in the

629
00:28:59,000 --> 00:29:03,039
back of my brain I can't think of the

630
00:29:00,960 --> 00:29:06,600
case but it's entirely it's entirely

631
00:29:03,039 --> 00:29:06,600
reasonable and possible that that will

632
00:29:07,279 --> 00:29:12,720
happen hi you mentioned a few potential

633
00:29:10,200 --> 00:29:15,200
future options on paths like Junctions

634
00:29:12,720 --> 00:29:16,919
and windows en codings is that something

635
00:29:15,200 --> 00:29:18,559
you start off the conversation with with

636
00:29:16,919 --> 00:29:20,240
the code owner of hey I've got a bunch

637
00:29:18,559 --> 00:29:21,760
of ideas and this is a prerequisite or

638
00:29:20,240 --> 00:29:25,000
do you just take it like one step at a

639
00:29:21,760 --> 00:29:29,440
time and see how far you get um I I

640
00:29:25,000 --> 00:29:31,519
would I would I would give B probably

641
00:29:29,440 --> 00:29:33,679
like this talk video basically so I'd

642
00:29:31,519 --> 00:29:34,840
give him the code and and the talk and

643
00:29:33,679 --> 00:29:36,679
and I'd sort of suggest you know these

644
00:29:34,840 --> 00:29:41,880
are all things that it might improve

645
00:29:36,679 --> 00:29:43,720
things um and you know uh bnie will have

646
00:29:41,880 --> 00:29:48,600
a much better idea of what the right way

647
00:29:43,720 --> 00:29:52,640
to go is for these things um you know I

648
00:29:48,600 --> 00:29:54,240
like none of this stuff is

649
00:29:52,640 --> 00:29:56,640
really

650
00:29:54,240 --> 00:29:58,360
wow all of these things are possible

651
00:29:56,640 --> 00:30:00,320
with python now it's just just that you

652
00:29:58,360 --> 00:30:03,399
got to do explicit code handling in your

653
00:30:00,320 --> 00:30:05,799
paths and around your path handling um

654
00:30:03,399 --> 00:30:07,600
and I think python could be nicer if

655
00:30:05,799 --> 00:30:09,480
some of that stuff was automated using

656
00:30:07,600 --> 00:30:12,840
the

657
00:30:09,480 --> 00:30:16,519
contexts thank you very much Clinton um

658
00:30:12,840 --> 00:30:20,519
I present to you the

659
00:30:16,519 --> 00:30:20,519
mug thank