> On 07/23/2015 02:09 PM, Raphael S Carvalho via Syslinux wrote: > >> > > My sincere opinion is to apply this patchset as-is, and incrementally > > improve multifs. Lack of alternatives (additional features) *should not* be > > a reason to block this patchset. Again, I really think that this patchset > > should be applied unless a technical reason, e.g. some deficiency > > introduced by one of the patches, says otherwise. HPA, what do you think? > > > > Step 1 is to commit it on a branch. > > -hpa >Without _a lot_ of testing, applying the patch "as-is" is not the best first step, IMHO. A separated branch sounds at least more reasonable in comparison. Additionally, the patch set seems to affect more-than-just-an-isolated file / feature, which means that incorporating other patches (still pending, or future ones) based on the current stable 6.03 would be more difficult, generally speaking. One important matter to consider is the syntax (i.e. users). A future discussion / email would sound something as "we already introduced a certain syntax for 'hd' and 'partitions', so now we are having troubles with supporting additional nomenclatures". Examples are: using space character(s) as alternative to the suggested comma (as chain.c32), or using labels and/or UUIDs (as chain.c32). Once a certain syntax is established for final users, improvements are more difficult if they are not considered in advance. I understand the desire of adding the multifs feature. I fear that (other) pending issues / regressions will be kept alive "forever". BTW, having a branch with this patch set "as-is" would mean propagating (even more) the unnecessary "fancy" forms of the term, instead of plain "multifs". Please, please, KISS, "multifs". Regards, Ady.> > _______________________________________________ > Syslinux mailing list > Submissions to Syslinux at zytor.com > Unsubscribe or set options at: > http://www.zytor.com/mailman/listinfo/syslinux >
Paulo Alcantara
2015-Aug-10 14:45 UTC
[syslinux] [PULL 0/8] MultiFS suppport for BIOS and EFI
Ady, Peter, et al. On Fri, July 24, 2015 5:28 pm, Ady via Syslinux wrote:> >> On 07/23/2015 02:09 PM, Raphael S Carvalho via Syslinux wrote: >> >> >> > My sincere opinion is to apply this patchset as-is, and incrementally >> > improve multifs. Lack of alternatives (additional features) *should >> not* be >> > a reason to block this patchset. Again, I really think that this >> patchset >> > should be applied unless a technical reason, e.g. some deficiency >> > introduced by one of the patches, says otherwise. HPA, what do you >> think? >> > >> >> Step 1 is to commit it on a branch. >> >> -hpa >> > > Without _a lot_ of testing, applying the patch "as-is" is not the best > first step, IMHO. A separated branch sounds at least more reasonable in > comparison. > > Additionally, the patch set seems to affect more-than-just-an-isolated > file / feature, which means that incorporating other patches (still > pending, or future ones) based on the current stable 6.03 would be more > difficult, generally speaking. > > One important matter to consider is the syntax (i.e. users). A future > discussion / email would sound something as "we already introduced a > certain syntax for 'hd' and 'partitions', so now we are having troubles > with supporting additional nomenclatures". Examples are: using space > character(s) as alternative to the suggested comma (as chain.c32), or > using labels and/or UUIDs (as chain.c32). Once a certain syntax is > established for final users, improvements are more difficult if they > are not considered in advance. > > I understand the desire of adding the multifs feature. I fear that > (other) pending issues / regressions will be kept alive "forever". > > BTW, having a branch with this patch set "as-is" would mean propagating > (even more) the unnecessary "fancy" forms of the term, instead of plain > "multifs". Please, please, KISS, "multifs".Raphael and I we're planning to integrate syntax used by chain.c32 module to support UUIDs, labels, MBR ids, GPT partition's GUIDs, etc. as we get some free time to do so. We've already pushed out a few patches on top this one that fix EFI support to correctly handle partitions outside the disk where Syslinux was installed in, and renaming of "MultiFS" to "multifs". There's also a branch where all multifs work is being done at: git://git.zytor.com/users/pcacjr.syslinux multifs-for-upstream (So, please, ignore this pull request) Thanks, Paulo -- Paulo Alcantara, C.E.S.A.R Speaking for myself only.
Paulo Alcantara
2015-Aug-10 14:52 UTC
[syslinux] [PULL 0/8] MultiFS suppport for BIOS and EFI
On Mon, August 10, 2015 11:45 am, Paulo Alcantara wrote:> Ady, Peter, et al. > > On Fri, July 24, 2015 5:28 pm, Ady via Syslinux wrote: >> >>> On 07/23/2015 02:09 PM, Raphael S Carvalho via Syslinux wrote: >>> >> >>> > My sincere opinion is to apply this patchset as-is, and incrementally >>> > improve multifs. Lack of alternatives (additional features) *should >>> not* be >>> > a reason to block this patchset. Again, I really think that this >>> patchset >>> > should be applied unless a technical reason, e.g. some deficiency >>> > introduced by one of the patches, says otherwise. HPA, what do you >>> think? >>> > >>> >>> Step 1 is to commit it on a branch. >>> >>> -hpa >>> >> >> Without _a lot_ of testing, applying the patch "as-is" is not the best >> first step, IMHO. A separated branch sounds at least more reasonable in >> comparison. >> >> Additionally, the patch set seems to affect more-than-just-an-isolated >> file / feature, which means that incorporating other patches (still >> pending, or future ones) based on the current stable 6.03 would be more >> difficult, generally speaking. >> >> One important matter to consider is the syntax (i.e. users). A future >> discussion / email would sound something as "we already introduced a >> certain syntax for 'hd' and 'partitions', so now we are having troubles >> with supporting additional nomenclatures". Examples are: using space >> character(s) as alternative to the suggested comma (as chain.c32), or >> using labels and/or UUIDs (as chain.c32). Once a certain syntax is >> established for final users, improvements are more difficult if they >> are not considered in advance. >> >> I understand the desire of adding the multifs feature. I fear that >> (other) pending issues / regressions will be kept alive "forever". >> >> BTW, having a branch with this patch set "as-is" would mean propagating >> (even more) the unnecessary "fancy" forms of the term, instead of plain >> "multifs". Please, please, KISS, "multifs". > > Raphael and I we're planning to integrate syntax used by chain.c32 module > to support UUIDs, labels, MBR ids, GPT partition's GUIDs, etc. as we get > some free time to do so. > > We've already pushed out a few patches on top this one that fix EFI > support to correctly handle partitions outside the disk where Syslinux was > installed in, and renaming of "MultiFS" to "multifs". > > There's also a branch where all multifs work is being done at: > git://git.zytor.com/users/pcacjr.syslinux multifs-for-upstreamThe repo is actually: git://git.zytor.com/users/pcacjr/syslinux.git (sorry for the typo :-) ) Paulo -- Paulo Alcantara, C.E.S.A.R Speaking for myself only.