@@ -1246,70 +1246,6 @@ of ``title_of('tutorial.txt')`` is exactly equal to the old one
1246
1246
stopped the build process from having to recompute
1247
1247
any of the consequences downstream of it.
1248
1248
1249
- Subtlety
1250
- ========
1251
-
1252
- <!-- HMM. WAIT. By this point we are kind of set.
1253
- The story is nice and seems done.
1254
- Do we really need a secion on how the graph can
1255
- change even as you traverse it?
1256
- I am not sure. Maybe it should be axed.
1257
- What do you think, Dan?
1258
- Am leaving it hear til you give an opinion. -->
1259
-
1260
-
1261
- In fact:
1262
-
1263
- ----------------- ignore everything -----------------
1264
- --------------- from here down -------------
1265
- --------------- unless dan chooses to resurrect it ------------
1266
- ----------- otherwise, we axe it in favor of a triumpant conclusion! --------
1267
-
1268
-
1269
-
1270
-
1271
- Why would the graph change as we are calculating it?
1272
-
1273
- The edges we have considered so far between documents are the result of
1274
- static site design — here, the fact that each HTML page has a link to
1275
- the preceding blog post. But sometimes edges arise from the content
1276
- itself! Blog posts, for example, might refer to each other
1277
- dynamically::
1278
-
1279
- I have been learning even more about the Pandas library.
1280
- You can read about my first steps in using it by visiting
1281
- my original `learning-pandas`_ blog post from last year.
1282
-
1283
- When this paragraph is rendered the output should look like:
1284
-
1285
- .. x...original `Learning About Pandas`_ blog post from last year.
1286
-
1287
- Therefore this HTML will need to be regenerated every time the title in
1288
- ``learning-pandas.rst `` is edited and changed.
1289
-
1290
- After running a rebuild step for a consequence, therefore, we will need
1291
- to rebuild the edges leading to it so that they reflect exactly the
1292
- inputs it in fact used during its rebuild. In the rare case that the
1293
- new set of edges includes one from a yet-to-be-rebuilt consequence
1294
- further along in the current topological sort, this will correctly
1295
- assure that the consequence then reappears in the ``todo() `` set. A
1296
- full replacement of all incoming edges is offered through a dedicated
1297
- graph method. If an update were added to the text of post A to mention
1298
- the later post C, then an edge would need to be generated to capture
1299
- that:
1300
-
1301
- x>>> g.add_edge('C.title', 'A.html')
1302
-
1303
- Thanks to this new edge, post A will now be considered one of
1304
- consequences of a change to the title of post C.
1305
-
1306
- x>>> g.recursive_consequences_of(['C.title'])
1307
- ['A.html']
1308
-
1309
- How can this mechanism be connected to actual code that takes the
1310
- current value of each node and builds the resulting consequences?
1311
- Python gives us many possible approaches. [Show various ways of
1312
- registering routines?]
1313
1249
1314
1250
1315
1251
.. Links
0 commit comments