Re: [DISCUSS] The Future of MD5Checker

2015-06-09 Thread Alex Harui
In fact, the SDK build broke downloading Saxon from some other server just the other day. Every once in a while, you aren’t going to get the bits you expected. And knowing that with an error message in the UI is better than a crash. I would prefer not to have us responsible for hosting the lates

Re: [DISCUSS] The Future of MD5Checker

2015-06-04 Thread Tom Chiverton
On 04/06/15 12:43, Nicholas Kwiatkowski wrote: Putting more in control of Apache I think would be a good idea... I'm less secure in the though that Adobe won't randomly move, remove or do weird things with the downloads we depend on. This would break the automated "copy the new version to Apach

Re: [DISCUSS] The Future of MD5Checker

2015-06-04 Thread Nicholas Kwiatkowski
The problems with the installer should be coming along soon. The HTTPS library in the Flash Player used the IE proxy on Windows based machines, which for users on older XP machines meant that HTTPS downloads from certain servers (those that use TLS only, and/or those that have newer G5 certs from V

Re: [DISCUSS] The Future of MD5Checker

2015-06-04 Thread Tom Chiverton
On 04/06/15 10:39, Kessler CTR Mark J wrote: They wouldn't be able to change until we changed them. So the MD5 would still match. Doesn't the installer have issues with downloading from Apache servers anyway, where the mirrors are broken ? So moving more stuff to them seems a bad idea ? T

RE: [DISCUSS] The Future of MD5Checker

2015-06-04 Thread Kessler CTR Mark J
They wouldn't be able to change until we changed them. So the MD5 would still match. -Original Message- From: Alex Harui [mailto:aha...@adobe.com] Sent: Wednesday, June 03, 2015 8:59 AM To: dev@flex.apache.org Subject: Re: [DISCUSS] The Future of MD5Checker On 6/3/15, 4:

Re: [DISCUSS] The Future of MD5Checker

2015-06-04 Thread Bertrand Delacretaz
Hi, On Wed, Jun 3, 2015 at 6:20 AM, Alex Harui wrote: > ...most folks manually download and unzip > a package and if that zip blows up, they just download again... Note that the main reason Apache provides digests alongside the files that we distribute is not to protect people against failed dow

Re: [DISCUSS] The Future of MD5Checker

2015-06-03 Thread Alex Harui
On 6/3/15, 4:04 AM, "Kessler CTR Mark J" wrote: >Side bar in the same context. Are we able to get permission to re-host >some of the files on apache servers since we are enforcing the users to >agree to the legal aspects(licenses / terms of usage) anyways? I haven’t tried for the Adobe stuff.

RE: [DISCUSS] The Future of MD5Checker

2015-06-03 Thread Kessler CTR Mark J
Side bar in the same context. Are we able to get permission to re-host some of the files on apache servers since we are enforcing the users to agree to the legal aspects(licenses / terms of usage) anyways? -Mark

Re: [DISCUSS] The Future of MD5Checker

2015-06-02 Thread OmPrakash Muppirala
I vote for 3 as well. As long as we have direct access to the VM and dont have any dependency on the Infra team. In my past several experiences dealing with the Infra team, the average wait time to get anything done is a few days. We cannot afford to go back to that. Thanks, Om On Jun 2, 201

Re: [DISCUSS] The Future of MD5Checker

2015-06-02 Thread Justin Mclean
Hi, > 1) Redirect to pull the list directly off my CI server. That would be 44K > per install (actually 88K since it gets pulled once by the Installer UI > and again by the Ant script. That might turn out to be more bandwidth > than I would like to support. > 2) Do #1, but move the MD5Checker ta

Re: [DISCUSS] The Future of MD5Checker

2015-06-02 Thread Alex Harui
On 6/2/15, 6:47 PM, "Chris Martin" wrote: >Not too familiar with the provenance of the MD5Checker or why we have to >generate our own hashes, so please forgive me if this is a “already >known" question 😊 Does adobe and Google host the expected MD5 hashes for >the artifacts that we download from

Re: [DISCUSS] The Future of MD5Checker

2015-06-02 Thread Chris Martin
Not too familiar with the provenance of the MD5Checker or why we have to generate our own hashes, so please forgive me if this is a “already known" question 😊 Does adobe and Google host the expected MD5 hashes for the artifacts that we download from them? If so, would it be possible to check aga