I received an issue about Boost PP having lots of old branches which the issuer would like removed. I notice most other Boost libraries also have lots of old branches. Is there any policy about removing these old branches from a Boost library repository ? Many of these old branches are svn-branches, but most of them do not have any value any more and, of course, do not affect the current 'develop' or 'master' branch.
Am 15.11.19 um 11:56 schrieb Edward Diener via Boost:
I received an issue about Boost PP having lots of old branches which the issuer would like removed. I notice most other Boost libraries also have lots of old branches. Is there any policy about removing these old branches from a Boost library repository ? Many of these old branches are svn-branches, but most of them do not have any value any more and, of course, do not affect the current 'develop' or 'master' branch.
It was me requesting this. See e.g. https://github.com/boostorg/preprocessor/branches/stale for about 70 stale branches. According to GitHub all(or at least most) of them are merged into master/develop as indicated by e.g. "1 commit ahead, 275 commits behind develop" where the "1 commit" is "This commit was manufactured by cvs2svn to create branch 'SPIRIT_1_6'." I don't see any value in those branches, as they look like feature branches from SVN times that are already merged and hence obsolete. Every Fork will have them too which makes switching to new feature branches when developing cumbersome. IMO one can delete all branches saying "1 commit ahead, *" where the latest commit is from "nobody" without further checking. The remaining ones might need a check if they actually contain new stuff.
On 11/15/2019 7:32 AM, Alexander Grund via Boost wrote:
Am 15.11.19 um 11:56 schrieb Edward Diener via Boost:
I received an issue about Boost PP having lots of old branches which the issuer would like removed. I notice most other Boost libraries also have lots of old branches. Is there any policy about removing these old branches from a Boost library repository ? Many of these old branches are svn-branches, but most of them do not have any value any more and, of course, do not affect the current 'develop' or 'master' branch.
It was me requesting this. See e.g. https://github.com/boostorg/preprocessor/branches/stale for about 70 stale branches. According to GitHub all(or at least most) of them are merged into master/develop as indicated by e.g. "1 commit ahead, 275 commits behind develop" where the "1 commit" is "This commit was manufactured by cvs2svn to create branch 'SPIRIT_1_6'."
I don't see any value in those branches, as they look like feature branches from SVN times that are already merged and hence obsolete. Every Fork will have them too which makes switching to new feature branches when developing cumbersome.
IMO one can delete all branches saying "1 commit ahead, *" where the latest commit is from "nobody" without further checking. The remaining ones might need a check if they actually contain new stuff.
I was just pointing out that Boost PP is hardly the only Boost library which has lots of stale branches. If it is not against Boost policy to delete old, stale branches I will be glad to do so, but I was just looking for some clarification from other Boost developers first.
Private reply intentionally: Completely understood your intention. Just wanted to give some background on the type of branches in question so decision makers have it easier. Am 15.11.19 um 13:49 schrieb Edward Diener via Boost:
On 11/15/2019 7:32 AM, Alexander Grund via Boost wrote:
Am 15.11.19 um 11:56 schrieb Edward Diener via Boost:
I received an issue about Boost PP having lots of old branches which the issuer would like removed. I notice most other Boost libraries also have lots of old branches. Is there any policy about removing these old branches from a Boost library repository ? Many of these old branches are svn-branches, but most of them do not have any value any more and, of course, do not affect the current 'develop' or 'master' branch.
It was me requesting this. See e.g. https://github.com/boostorg/preprocessor/branches/stale for about 70 stale branches. According to GitHub all(or at least most) of them are merged into master/develop as indicated by e.g. "1 commit ahead, 275 commits behind develop" where the "1 commit" is "This commit was manufactured by cvs2svn to create branch 'SPIRIT_1_6'."
I don't see any value in those branches, as they look like feature branches from SVN times that are already merged and hence obsolete. Every Fork will have them too which makes switching to new feature branches when developing cumbersome.
IMO one can delete all branches saying "1 commit ahead, *" where the latest commit is from "nobody" without further checking. The remaining ones might need a check if they actually contain new stuff.
I was just pointing out that Boost PP is hardly the only Boost library which has lots of stale branches. If it is not against Boost policy to delete old, stale branches I will be glad to do so, but I was just looking for some clarification from other Boost developers first.
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Alexander Grund Interdisziplinäre Anwendungsunterstützung und Koordination (IAK) Technische Universität Dresden Zentrum für Informationsdienste und Hochleistungsrechnen (ZIH) Würzburger Str.35/Chemnitzer Str.50, Raum 010 01062 Dresden Tel.: +49 (351) 463-35982 E-Mail: alexander.grund@tu-dresden.de ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Edward Diener wrote:
I was just pointing out that Boost PP is hardly the only Boost library which has lots of stale branches. If it is not against Boost policy to delete old, stale branches I will be glad to do so, but I was just looking for some clarification from other Boost developers first.
It is not against any policy. A maintainer is free to delete old branches. Only develop and master need to be around. Glen
On 11/15/2019 7:51 AM, Alexander Grund via Boost wrote:
Private reply intentionally:
Completely understood your intention. Just wanted to give some background on the type of branches in question so decision makers have it easier.
I removed all the stale branches in Boost PP.
Am 15.11.19 um 13:49 schrieb Edward Diener via Boost:
On 11/15/2019 7:32 AM, Alexander Grund via Boost wrote:
Am 15.11.19 um 11:56 schrieb Edward Diener via Boost:
I received an issue about Boost PP having lots of old branches which the issuer would like removed. I notice most other Boost libraries also have lots of old branches. Is there any policy about removing these old branches from a Boost library repository ? Many of these old branches are svn-branches, but most of them do not have any value any more and, of course, do not affect the current 'develop' or 'master' branch.
It was me requesting this. See e.g. https://github.com/boostorg/preprocessor/branches/stale for about 70 stale branches. According to GitHub all(or at least most) of them are merged into master/develop as indicated by e.g. "1 commit ahead, 275 commits behind develop" where the "1 commit" is "This commit was manufactured by cvs2svn to create branch 'SPIRIT_1_6'."
I don't see any value in those branches, as they look like feature branches from SVN times that are already merged and hence obsolete. Every Fork will have them too which makes switching to new feature branches when developing cumbersome.
IMO one can delete all branches saying "1 commit ahead, *" where the latest commit is from "nobody" without further checking. The remaining ones might need a check if they actually contain new stuff.
I was just pointing out that Boost PP is hardly the only Boost library which has lots of stale branches. If it is not against Boost policy to delete old, stale branches I will be glad to do so, but I was just looking for some clarification from other Boost developers first.
On 2019-11-15 15:49, Edward Diener via Boost wrote:
On 11/15/2019 7:32 AM, Alexander Grund via Boost wrote:
Am 15.11.19 um 11:56 schrieb Edward Diener via Boost:
I received an issue about Boost PP having lots of old branches which the issuer would like removed. I notice most other Boost libraries also have lots of old branches. Is there any policy about removing these old branches from a Boost library repository ? Many of these old branches are svn-branches, but most of them do not have any value any more and, of course, do not affect the current 'develop' or 'master' branch.
It was me requesting this. See e.g. https://github.com/boostorg/preprocessor/branches/stale for about 70 stale branches. According to GitHub all(or at least most) of them are merged into master/develop as indicated by e.g. "1 commit ahead, 275 commits behind develop" where the "1 commit" is "This commit was manufactured by cvs2svn to create branch 'SPIRIT_1_6'."
I don't see any value in those branches, as they look like feature branches from SVN times that are already merged and hence obsolete. Every Fork will have them too which makes switching to new feature branches when developing cumbersome.
IMO one can delete all branches saying "1 commit ahead, *" where the latest commit is from "nobody" without further checking. The remaining ones might need a check if they actually contain new stuff.
I was just pointing out that Boost PP is hardly the only Boost library which has lots of stale branches. If it is not against Boost policy to delete old, stale branches I will be glad to do so, but I was just looking for some clarification from other Boost developers first.
I believe, beyond master and develop, your usage of branches is completely at your (project members') discretion. There's no Boost-wide policy.
I was just pointing out that Boost PP is hardly the only Boost library which has lots of stale branches. If it is not against Boost policy to delete old, stale branches I will be glad to do so, but I was just looking for some clarification from other Boost developers first.
I believe, beyond master and develop, your usage of branches is completely at your (project members') discretion. There's no Boost-wide policy.
It is surely sensible that fully merged branches be deleted as a matter of default? I very rarely keep around fully merged branches, but I do occasionally do so mainly to retain a waypoint reference. Niall
participants (5)
-
Alexander Grund -
Andrey Semashev -
Edward Diener -
Glen Fernandes -
Niall Douglas