openstack
/
stevedore
Code Issues Proposed changes
stevedore/8eb97446db69bc34fee40b5a06a...

218 lines
10 KiB
Plaintext
Raw Blame History

{
"comments": [
{
"unresolved": false,
"key": {
"uuid": "0d9f379f_41c2da57",
"filename": "/PATCHSET_LEVEL",
"patchSetId": 1
},
"lineNbr": 0,
"author": {
"id": 28522
},
"writtenOn": "2022-09-09T16:16:46Z",
"side": 1,
"message": "Thanks for your detailed response",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "fc883b6b_22b69b3e",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 15334
},
"writtenOn": "2022-09-09T15:57:43Z",
"side": 1,
"message": "I don\u0027t know if I should be doing this. As things stand, these brackets aren\u0027t currently present on Python 3.10 or 3.9.11, only 3.9.10 or lower. We\u0027re changing behavior here and forcing ourselves to carry this forward forever. However, if we don\u0027t add these then we\u0027ll see different results for Python 3.7 and 3.8 unless we change behavior there also.\n\nI\u0027d be tempted to *not* add this and change the behavior for all Python versions (with a release note and potentially major version bump) but I\u0027d like others opinions.",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "21aa3d83_759d4bca",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 28522
},
"writtenOn": "2022-09-09T16:05:12Z",
"side": 1,
"message": "You propose to change the behavior for all the python versions by abandoning the current solution, or you means that this patch will change the behavior for all Python versions?",
"parentUuid": "fc883b6b_22b69b3e",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "32312b15_e4782530",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 15334
},
"writtenOn": "2022-09-09T16:12:14Z",
"side": 1,
"message": "Currently:\n\n Python 3.7.x -\u003e [entry]\n Python 3.8.x -\u003e [entry]\n Python 3.9.x (x \u003c 11) -\u003e [entry]\n Python 3.9.x (x \u003e\u003d 11) -\u003e entry\n Python 3.10.x -\u003e entry\n\nOption A:\n\n Python 3.7.x -\u003e [entry]\n Python 3.8.x -\u003e [entry]\n Python 3.9.x (x \u003c 11) -\u003e [entry]\n Python 3.9.x (x \u003e\u003d 11) -\u003e [entry]\n Python 3.10.x -\u003e [entry]\n\nOption B:\n\n Python 3.7.x -\u003e entry\n Python 3.8.x -\u003e entry\n Python 3.9.x (x \u003c 11) -\u003e entry\n Python 3.9.x (x \u003e\u003d 11) -\u003e entry\n Python 3.10.x -\u003e entry\n\nEither way we\u0027re potentially breaking someone. Question is who.\n\nIf it helps, I can\u0027t find any references to this function via codesearch.o.o. Doesn\u0027t mean someone else isn\u0027t relying on it though...",
"parentUuid": "21aa3d83_759d4bca",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "6643c120_7c4b2922",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 15334
},
"writtenOn": "2022-09-09T16:12:56Z",
"side": 1,
"message": "s/entry/extra/ for all those rows above /o\\ You get the idea",
"parentUuid": "32312b15_e4782530",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "e2ffea3d_faaa5643",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 28522
},
"writtenOn": "2022-09-09T16:16:46Z",
"side": 1,
"message": "I see, then I\u0027d suggest to break the oldest versions of Python (3.7, 3.8, 3.9), as the other are the future of our daily usages.",
"parentUuid": "6643c120_7c4b2922",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "1041ac9c_83a4af5e",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 28522
},
"writtenOn": "2022-09-09T16:17:12Z",
"side": 1,
"message": "+1 for the option B",
"parentUuid": "e2ffea3d_faaa5643",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "6b264bc8_1ecbb2da",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 28522
},
"writtenOn": "2022-09-09T16:20:56Z",
"side": 1,
"message": "Because, apart from the fact that that we should remind us to remove this fix once we stop supporting version of Python \u003c 3.10, we won\u0027t have to adapt something else.",
"parentUuid": "1041ac9c_83a4af5e",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": false,
"key": {
"uuid": "9909e1fc_7ba2c2a4",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 15334
},
"writtenOn": "2022-10-03T16:24:59Z",
"side": 1,
"message": "Done",
"parentUuid": "6b264bc8_1ecbb2da",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "a99c6dee_6c6d62c7",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 2472
},
"writtenOn": "2022-10-03T20:20:53Z",
"side": 1,
"message": "Are there any tests for this behavior to demonstrate what might break?\n\nDo we know of any consumers of extras?\n\nThe packaging docs imply this feature is deprecated. Maybe we remove the extras property implementation entirely? Would that break OpenStack?\n\nhttps://packaging.python.org/en/latest/specifications/entry-points/#data-model",
"parentUuid": "21aa3d83_759d4bca",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "1127fd4d_28a23b61",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 15334
},
"writtenOn": "2022-10-04T10:22:48Z",
"side": 1,
"message": "\u003e Are there any tests for this behavior to demonstrate what might break?\n\nYou\u0027ll note I had to change a test in this change. This is the only thing that should change. If there were users of this that were either relying on the brackets being present or were stripping these brackets in a non-robust way (i.e. things crash if the brackets aren\u0027t present), they would be broken by this. However, with the current state of things they\u0027re already going to be broken by Python 3.9.11 or 3.12.x onwards, so not a whole lot is changing here.\n\n\u003e Do we know of any consumers of extras?\n\nNo, I couldn\u0027t find any via codesearch.o.o\n\n\u003e The packaging docs imply this feature is deprecated. Maybe we remove the extras property implementation entirely? Would that break OpenStack?\n\u003e \n\u003e https://packaging.python.org/en/latest/specifications/entry-points/#data-model\n\nGood idea. There are no OpenStack users so we don\u0027t have an issue there. Such a change would not be backportable to stable/zed though. This is an issue there since we support Python 3.9 in Zed, and 3.9.11 includes a backported fix to importlib.metadata that introduces this behavior change. Perhaps I should submit this in its previous incarnation (retaining the brackets) and then do a follow-up to remove the feature entirely, followed by a major version bump?",
"parentUuid": "a99c6dee_6c6d62c7",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
},
{
"unresolved": true,
"key": {
"uuid": "b9025750_18581847",
"filename": "stevedore/extension.py",
"patchSetId": 1
},
"lineNbr": 75,
"author": {
"id": 2472
},
"writtenOn": "2022-10-04T13:35:43Z",
"side": 1,
"message": "\u003e \u003e Are there any tests for this behavior to demonstrate what might break?\n\u003e \n\u003e You\u0027ll note I had to change a test in this change. \n\nI missed that the first time through.\n\n\u003e \u003e The packaging docs imply this feature is deprecated. Maybe we remove the extras property implementation entirely? Would that break OpenStack?\n\u003e \u003e \n\u003e \u003e https://packaging.python.org/en/latest/specifications/entry-points/#data-model\n\u003e \n\u003e Good idea. There are no OpenStack users so we don\u0027t have an issue there. Such a change would not be backportable to stable/zed though. This is an issue there since we support Python 3.9 in Zed, and 3.9.11 includes a backported fix to importlib.metadata that introduces this behavior change. Perhaps I should submit this in its previous incarnation (retaining the brackets) and then do a follow-up to remove the feature entirely, followed by a major version bump?\n\nI think I like that approach. It\u0027s a cleaner break, and since the feature is deprecated we\u0027re likely to have to do it in the future anyway.",
"parentUuid": "1127fd4d_28a23b61",
"revId": "8eb97446db69bc34fee40b5a06ac0a9156e204f4",
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543"
}
]
}
View Git Blame Copy Permalink