{"thread":{"id":"37801","subject":"Re: [PATCH] Expand on impact of threads on memory consumption.","startedAt":"2014-10-24T17:51:15Z","lastAt":"2014-10-28T16:57:50Z","messageCount":2,"participants":["Robert de Bath","Junio C Hamano"],"isPatch":true,"patchVersion":1,"patchTotal":null},"messages":[{"id":"251024","messageId":"3F70E0BBE32F764BA89060B84135BD471850D9E1@TVTEX10.tvt.local","threadId":"37801","inReplyTo":null,"subject":"Re: [PATCH] Expand on impact of threads on memory consumption.","fromName":"Robert de Bath","fromEmail":"robert@tvisiontech.co.uk","sentAt":"2014-10-24T17:51:15Z","receivedAt":"2014-10-24T17:51:15Z","isPatch":true,"sender":{"key":"robert@tvisiontech.co.uk","avatar":null},"body":"> ... text unnecessarily opaque ...\nYes, that's quite likely. \nHowever, I think you should continue to have a similar form to the previous 'pack.window' and \n'pack.depth' descriptions... perhaps:\n\n The maximum size of memory that is consumed by each\n thread in linkgit:git-pack-objects[1] for pack window memory\n when no limit is given on the command line.\n The value can be suffixed with \"k\", \"m\", or \"g\". When left\n unconfigured, there will be no limit.\n\nThough, I suspect the existing wording is trying to avoid saying there's a 'default' and \nanother 'default' which would be covered by your rewording.\n\nSo I have no real problem either way.\n\nRobert.\n \n________________________________________\nFrom: Junio C Hamano [gitster@pobox.com]\nSent: 24 October 2014 18:27\nTo: Robert de Bath\nCc: git@vger.kernel.org\nSubject: [NOTWHITELIST] Re: [PATCH] Expand on impact of threads on memory consumption.\n\nRobert de Bath writes:\n\n> diff --git a/Documentation/config.txt b/Documentation/config.txt\n> index c55c22a..6384038 100644\n> --- a/Documentation/config.txt\n> +++ b/Documentation/config.txt\n> @@ -1828,7 +1828,10 @@ pack.depth::\n>\n> pack.windowMemory::\n> The window memory size limit used by linkgit:git-pack-objects[1]\n> - when no limit is given on the command line. The value can be\n> + when no limit is given on the command line. This limit may be\n> + allocated individually by each thread doing delta compression\n> + so the total memory allocated by linkgit:git-pack-objects[1] can\n> + be this multiplied by the number of threads. The value can be\n> suffixed with \"k\", \"m\", or \"g\". Defaults to 0, meaning no\n> limit.\n\nI sense that this change makes the resulting text unnecessarily\nopaque by avoiding touching the first sentence. Wouldn't something\nmore concise like this sufficient?\n\n The default maximum size of memory that is consumed by each\n thread in linkgit:git-pack-objects[1] for pack window memory.\n The value can be suffixed with \"k\", \"m\", or \"g\". When left\n unconfigured, there will be no limit.\n"},{"id":"251146","messageId":"xmqqvbn43ze9.fsf@gitster.dls.corp.google.com","threadId":"37801","inReplyTo":"3F70E0BBE32F764BA89060B84135BD471850D9E1@TVTEX10.tvt.local","subject":"Re: [PATCH] Expand on impact of threads on memory consumption.","fromName":"Junio C Hamano","fromEmail":"gitster@pobox.com","sentAt":"2014-10-28T16:57:50Z","receivedAt":"2014-10-28T16:57:50Z","isPatch":true,"sender":{"key":"gitster@pobox.com","avatar":"https://avatars.githubusercontent.com/u/54884?v=4"},"body":"Robert de Bath writes:\n\n>> ... text unnecessarily opaque ...\n> Yes, that's quite likely. \n> However, I think you should continue to have a similar form to the\n> previous 'pack.window' and\n> 'pack.depth' descriptions... perhaps:\n>\n> The maximum size of memory that is consumed by each\n> thread in linkgit:git-pack-objects[1] for pack window memory\n> when no limit is given on the command line.\n> The value can be suffixed with \"k\", \"m\", or \"g\". When left\n> unconfigured, there will be no limit.\n\nOK, let's go with that wording.\n\nThanks.\n\n>\n> Though, I suspect the existing wording is trying to avoid saying there's a 'default' and \n> another 'default' which would be covered by your rewording.\n>\n> So I have no real problem either way.\n>\n> Robert.\n> \n> ________________________________________\n> From: Junio C Hamano [gitster@pobox.com]\n> Sent: 24 October 2014 18:27\n> To: Robert de Bath\n> Cc: git@vger.kernel.org\n> Subject: [NOTWHITELIST] Re: [PATCH] Expand on impact of threads on memory consumption.\n>\n> Robert de Bath writes:\n>\n>> diff --git a/Documentation/config.txt b/Documentation/config.txt\n>> index c55c22a..6384038 100644\n>> --- a/Documentation/config.txt\n>> +++ b/Documentation/config.txt\n>> @@ -1828,7 +1828,10 @@ pack.depth::\n>>\n>> pack.windowMemory::\n>> The window memory size limit used by linkgit:git-pack-objects[1]\n>> - when no limit is given on the command line. The value can be\n>> + when no limit is given on the command line. This limit may be\n>> + allocated individually by each thread doing delta compression\n>> + so the total memory allocated by linkgit:git-pack-objects[1] can\n>> + be this multiplied by the number of threads. The value can be\n>> suffixed with \"k\", \"m\", or \"g\". Defaults to 0, meaning no\n>> limit.\n>\n> I sense that this change makes the resulting text unnecessarily\n> opaque by avoiding touching the first sentence. Wouldn't something\n> more concise like this sufficient?\n>\n> The default maximum size of memory that is consumed by each\n> thread in linkgit:git-pack-objects[1] for pack window memory.\n> The value can be suffixed with \"k\", \"m\", or \"g\". When left\n> unconfigured, there will be no limit.\n"}]}