<feed xmlns='http://www.w3.org/2005/Atom'>
<title>progmaweb/.github/workflows, branch site/phone-layout</title>
<subtitle>The site and account service of Progmasoft: a Next.js frontend and an ASP.NET Core API.</subtitle>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/'/>
<entry>
<title>Fit every page to phones and test the layout with Playwright</title>
<updated>2026-11-16T17:28:33+00:00</updated>
<author>
<name>Leitwolf11</name>
</author>
<published>2026-11-16T17:28:33+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=537b1f3a4a605b7aa980eec520628ff37648f683'/>
<id>537b1f3a4a605b7aa980eec520628ff37648f683</id>
<content type='text'>
On a phone the header hid every link but the account button, so the
organization site had no way to its products or developer pages, and the
ViGet, Account and Console headers wrapped onto two or three lines.

- HeaderMenu: below 980px the header holds the brand and a menu button;
  the button opens the links and the preferences as a panel under the
  header (48px rows, 44px controls). Escape, a chosen link or a second
  press closes it. Below 420px the button keeps only its icon, so a brand
  with a product name ("Console", "Account") is never covered. All three
  headers (site, ViGet, credentials) use it; ViGet's own column layout is
  gone. Labels in every locale.
- Header links on wide screens get 10px of vertical padding: the same look,
  40px targets.
- The consent checkbox is 24px.
- product.progmasoft.com: below 560px each project row is a card with
  labelled facts instead of a table that scrolled sideways.
- progmasoft.com: the area tiles beside the hero repeated the list right
  below it; they are hidden once the hero is one column. The OAuth sample
  wraps on phones instead of being cut off.

@playwright/test 1.65.0 runs the built site (next start) in Chromium at
360, 390, 768 and 1280px. Every public host name resolves to the local
server through --host-resolver-rules, so no request leaves the machine.
For one page of each kind on every host it checks: no horizontal scroll,
nothing past the screen edge outside a scrolling box, controls of at least
24 by 24px (WCAG 2.2, 2.5.8), the header on one line with brand and
controls not covering each other; and that the phone menu opens and closes.
The Frontend workflow installs Chromium and runs it after the build,
keeping the report and traces when it fails. 140 tests pass locally; with
the 420px rule removed the overlap check fails on Account and Console.

Claude-Session: https://claude.ai/code/session_01Ji7xZMYjc4zankoc7hGBE1
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
On a phone the header hid every link but the account button, so the
organization site had no way to its products or developer pages, and the
ViGet, Account and Console headers wrapped onto two or three lines.

- HeaderMenu: below 980px the header holds the brand and a menu button;
  the button opens the links and the preferences as a panel under the
  header (48px rows, 44px controls). Escape, a chosen link or a second
  press closes it. Below 420px the button keeps only its icon, so a brand
  with a product name ("Console", "Account") is never covered. All three
  headers (site, ViGet, credentials) use it; ViGet's own column layout is
  gone. Labels in every locale.
- Header links on wide screens get 10px of vertical padding: the same look,
  40px targets.
- The consent checkbox is 24px.
- product.progmasoft.com: below 560px each project row is a card with
  labelled facts instead of a table that scrolled sideways.
- progmasoft.com: the area tiles beside the hero repeated the list right
  below it; they are hidden once the hero is one column. The OAuth sample
  wraps on phones instead of being cut off.

@playwright/test 1.65.0 runs the built site (next start) in Chromium at
360, 390, 768 and 1280px. Every public host name resolves to the local
server through --host-resolver-rules, so no request leaves the machine.
For one page of each kind on every host it checks: no horizontal scroll,
nothing past the screen edge outside a scrolling box, controls of at least
24 by 24px (WCAG 2.2, 2.5.8), the header on one line with brand and
controls not covering each other; and that the phone menu opens and closes.
The Frontend workflow installs Chromium and runs it after the build,
keeping the report and traces when it fails. 140 tests pass locally; with
the 420px rule removed the overlap check fails on Account and Console.

Claude-Session: https://claude.ai/code/session_01Ji7xZMYjc4zankoc7hGBE1
</pre>
</div>
</content>
</entry>
<entry>
<title>Name the copyright license alone in SPDX expressions</title>
<updated>2026-11-15T10:14:38+00:00</updated>
<author>
<name>Leitwolf11</name>
</author>
<published>2026-11-15T10:14:38+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=3bda6d7bc28c5aa588418d12d9e3129f8526df50'/>
<id>3bda6d7bc28c5aa588418d12d9e3129f8526df50</id>
<content type='text'>
The Progmasoft Patent Grant is a separate permission added to the
copyright license, not an exception to it, so joining the two with WITH
said something the grant does not do. What makes the grant apply is the
PATENTS notice of the repository (Section 9 of the grant), not the SPDX
expression of each file.

Every project-owned file now carries SPDX-License-Identifier:
AGPL-3.0-or-later, and package.json says the same. PATENTS becomes
PATENTS.md: project-owned software in the repository is covered by the
Patent Grant, Version 1.1, whose canonical text is unchanged under
LICENSES/; third-party material keeps its own terms. NOTICE.txt and the
README point to PATENTS.md.

Claude-Session: https://claude.ai/code/session_01Ji7xZMYjc4zankoc7hGBE1
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
The Progmasoft Patent Grant is a separate permission added to the
copyright license, not an exception to it, so joining the two with WITH
said something the grant does not do. What makes the grant apply is the
PATENTS notice of the repository (Section 9 of the grant), not the SPDX
expression of each file.

Every project-owned file now carries SPDX-License-Identifier:
AGPL-3.0-or-later, and package.json says the same. PATENTS becomes
PATENTS.md: project-owned software in the repository is covered by the
Patent Grant, Version 1.1, whose canonical text is unchanged under
LICENSES/; third-party material keeps its own terms. NOTICE.txt and the
README point to PATENTS.md.

Claude-Session: https://claude.ai/code/session_01Ji7xZMYjc4zankoc7hGBE1
</pre>
</div>
</content>
</entry>
<entry>
<title>Take the pnpm version from packageManager only</title>
<updated>2026-11-13T17:25:35+00:00</updated>
<author>
<name>Leitwolf11</name>
</author>
<published>2026-11-13T17:25:35+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=6e706d4da3c95e913c1946af1507f98d01356271'/>
<id>6e706d4da3c95e913c1946af1507f98d01356271</id>
<content type='text'>
pnpm/action-setup v5 stops when the workflow names a pnpm version and
package.json names another in packageManager. The workflows no longer
name one, so packageManager (pnpm 11.24.0) is the only source, as it is
for every developer using Corepack.

Co-Authored-By: Claude Opus 5.5 &lt;noreply@anthropic.com&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
pnpm/action-setup v5 stops when the workflow names a pnpm version and
package.json names another in packageManager. The workflows no longer
name one, so packageManager (pnpm 11.24.0) is the only source, as it is
for every developer using Corepack.

Co-Authored-By: Claude Opus 5.5 &lt;noreply@anthropic.com&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>Update GitHub Actions</title>
<updated>2026-11-12T20:43:14+00:00</updated>
<author>
<name>renovate[bot]</name>
</author>
<published>2026-11-12T20:43:14+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=2f852a2487be5e61989daaca59afa287bbd2728c'/>
<id>2f852a2487be5e61989daaca59afa287bbd2728c</id>
<content type='text'>
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
</pre>
</div>
</content>
</entry>
<entry>
<title>Rate-limit sign-in by visitor behind Cloudflare and stop forwarding h2c</title>
<updated>2026-11-08T21:38:47+00:00</updated>
<author>
<name>Leitwolf11</name>
</author>
<published>2026-11-08T21:38:47+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=66f3cc196a9b91d471a90e021b12630100181c4a'/>
<id>66f3cc196a9b91d471a90e021b12630100181c4a</id>
<content type='text'>
Cloudflare proxies the Progmasoft hosts, so Nginx saw a Cloudflare edge as
its peer and appended that edge to X-Forwarded-For. The API takes the last
hop as the client, so the account-auth limit (12 attempts in 3 minutes) was
counted per edge: 38 failed sign-ins from one address in a row were all
answered with 401, none with 429, and unrelated visitors behind one edge
shared a bucket.

- ops/nginx/cloudflare-real-ip.conf trusts CF-Connecting-IP only from the
  published Cloudflare ranges; a direct peer keeps its own address, so the
  header cannot be forged around Cloudflare. With it installed the 13th
  failed sign-in gets 429, a forged X-Forwarded-For changes nothing, and
  another address is not affected.
- Every proxied location forwarded the client's Upgrade header with
  "Connection: upgrade" on each request. Only a WebSocket upgrade is now
  forwarded; other requests carry no Upgrade and an empty Connection.
- c8 reached glob 13 (GHSA-3g3m-2fjr-6pw4) and minimatch 10
  (GHSA-g3hv-75wg-m249) through test-exclude; both are overridden to the
  fixed majors. pnpm audit reports nothing, coverage, lint and check pass.
- The CodeQL action pins carry their exact version comment (v4.38.2).
- ops/README.md describes the address handling and no longer names a
  script that left the repository with AI Support.

Co-Authored-By: Claude Opus 5.5 &lt;noreply@anthropic.com&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Cloudflare proxies the Progmasoft hosts, so Nginx saw a Cloudflare edge as
its peer and appended that edge to X-Forwarded-For. The API takes the last
hop as the client, so the account-auth limit (12 attempts in 3 minutes) was
counted per edge: 38 failed sign-ins from one address in a row were all
answered with 401, none with 429, and unrelated visitors behind one edge
shared a bucket.

- ops/nginx/cloudflare-real-ip.conf trusts CF-Connecting-IP only from the
  published Cloudflare ranges; a direct peer keeps its own address, so the
  header cannot be forged around Cloudflare. With it installed the 13th
  failed sign-in gets 429, a forged X-Forwarded-For changes nothing, and
  another address is not affected.
- Every proxied location forwarded the client's Upgrade header with
  "Connection: upgrade" on each request. Only a WebSocket upgrade is now
  forwarded; other requests carry no Upgrade and an empty Connection.
- c8 reached glob 13 (GHSA-3g3m-2fjr-6pw4) and minimatch 10
  (GHSA-g3hv-75wg-m249) through test-exclude; both are overridden to the
  fixed majors. pnpm audit reports nothing, coverage, lint and check pass.
- The CodeQL action pins carry their exact version comment (v4.38.2).
- ops/README.md describes the address handling and no longer names a
  script that left the repository with AI Support.

Co-Authored-By: Claude Opus 5.5 &lt;noreply@anthropic.com&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>Merge main into the Renovate branch</title>
<updated>2026-10-10T20:16:50+00:00</updated>
<author>
<name>Leitwolf11</name>
</author>
<published>2026-10-10T20:16:50+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=107a257bb5313828ca5a36e4f1c2963dfb24b294'/>
<id>107a257bb5313828ca5a36e4f1c2963dfb24b294</id>
<content type='text'>
main gained the workflow analysis change, which adds a line to the
checkout step right above the action whose version comment this branch
corrects. Both changes are kept: the checkout does not persist its
credentials, and the comment names v4.9.0.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
main gained the workflow analysis change, which adds a line to the
checkout step right above the action whose version comment this branch
corrects. Both changes are kept: the checkout does not persist its
credentials, and the comment names v4.9.0.
</pre>
</div>
</content>
</entry>
<entry>
<title>Name the pinned version of the dependency review action</title>
<updated>2026-10-10T19:35:08+00:00</updated>
<author>
<name>Leitwolf11</name>
</author>
<published>2026-10-10T19:35:08+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=65f4b4f309958def71b9b8956c06d0d55b75e700'/>
<id>65f4b4f309958def71b9b8956c06d0d55b75e700</id>
<content type='text'>
The action is pinned to the commit of v4.9.0, and the comment beside
the pin said v4. Renovate reads the comment as the version in use and
looked for a tag named v5 to update to; the action has v5.0.0 and no
v5, so the lookup failed and Renovate warned about it. The comment now
names the version the commit is.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
The action is pinned to the commit of v4.9.0, and the comment beside
the pin said v4. Renovate reads the comment as the version in use and
looked for a tag named v5 to update to; the action has v5.0.0 and no
v5, so the lookup failed and Renovate warned about it. The comment now
names the version the commit is.
</pre>
</div>
</content>
</entry>
<entry>
<title>Leave out the Scorecard workflow</title>
<updated>2026-10-10T17:26:52+00:00</updated>
<author>
<name>Leitwolf11</name>
</author>
<published>2026-10-10T17:26:52+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=04622baa18a5d52b60ec0f7ea882896956a91c3a'/>
<id>04622baa18a5d52b60ec0f7ea882896956a91c3a</id>
<content type='text'>
The owner decided against running OpenSSF Scorecard here. The workflow
added by the commit before this one is removed; nothing else changes.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
The owner decided against running OpenSSF Scorecard here. The workflow
added by the commit before this one is removed; nothing else changes.
</pre>
</div>
</content>
</entry>
<entry>
<title>Add workflow and repository analysis</title>
<updated>2026-10-10T17:23:08+00:00</updated>
<author>
<name>Leitwolf11</name>
</author>
<published>2026-10-10T17:23:08+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=525f3c0566247bbb6a81b5a2e1488147688c089a'/>
<id>525f3c0566247bbb6a81b5a2e1488147688c089a</id>
<content type='text'>
zizmor audits the workflows themselves for what makes a workflow
unsafe, and uploads what it finds to code scanning. Its first run here
found one thing, 6 times: a checkout that leaves the job's token in
the checked-out repository, where any later step can read it. No
workflow pushes or needs the token after the checkout, so every checkout
now sets persist-credentials to false, and zizmor reports nothing.

OpenSSF Scorecard grades how the repository is kept: branch protection,
pinned dependencies, token permissions and the like. It runs on the
default branch and weekly, publishes its score and uploads its findings to
code scanning.

Every new action is pinned to a commit, each job names the permissions
it needs, and no new workflow has a default permission.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
zizmor audits the workflows themselves for what makes a workflow
unsafe, and uploads what it finds to code scanning. Its first run here
found one thing, 6 times: a checkout that leaves the job's token in
the checked-out repository, where any later step can read it. No
workflow pushes or needs the token after the checkout, so every checkout
now sets persist-credentials to false, and zizmor reports nothing.

OpenSSF Scorecard grades how the repository is kept: branch protection,
pinned dependencies, token permissions and the like. It runs on the
default branch and weekly, publishes its score and uploads its findings to
code scanning.

Every new action is pinned to a commit, each job names the permissions
it needs, and no new workflow has a default permission.
</pre>
</div>
</content>
</entry>
<entry>
<title>Keep accounts and sessions in PostgreSQL</title>
<updated>2026-10-06T19:53:46+00:00</updated>
<author>
<name>Leitwolf11</name>
</author>
<published>2026-10-06T19:53:46+00:00</published>
<link rel='alternate' type='text/html' href='https://git.progmasoft.com/progmaweb/commit/?id=1c986ff99b3c8f349e4a63c0e922276ddf7dc836'/>
<id>1c986ff99b3c8f349e4a63c0e922276ddf7dc836</id>
<content type='text'>
The account API kept everything in the memory of its process, so every
restart and every deployment of the API deleted all accounts and signed
everyone out.

PostgresAccountStore implements IAccountStore on PostgreSQL through Npgsql.
The three uniqueness rules are unique constraints, so two requests that race
for one name, address or Google identity are decided by the database. A
refused account is an ordinary answer, not an exception, and the rules are
reported in the order the in-memory store reports them. A session refers to
its account, and a session of an account that is not stored is refused; the
in-memory store now refuses it as well.

The schema is a list of numbered steps. The service takes the missing ones
when it starts, in one transaction under an advisory lock, and does not
start when the database cannot be reached or is newer than the release.
/health answers 503 while the database does not answer.

ConnectionStrings:Accounts selects PostgreSQL. A production service that is
given no database does not start; keeping accounts in memory there has to
be asked for with Accounts:Store=Memory. Development and the tests keep the
in-memory store by default.

The store gains RemoveExpiredSessionsAsync, and a background service calls
it at the start and once an hour: a session whose token is never presented
again no longer keeps its row for ever.

PostgreSQL keeps a timestamp to the microsecond. The store cuts a moment to
that precision and returns what it stored, and the service uses the
returned account, so the creation time in the registration response equals
the one every later read returns.

Tests. AccountStoreContractTests states the rules once and runs them on
both stores. Further tests cover the schema steps applied concurrently and a
database that is too new, the choice of the store from the configuration,
the removal of expired sessions, and an account with its session surviving
a restart of the service over HTTP. The PostgreSQL tests read their
database from PROGMAWEB_TEST_POSTGRES and work in a schema of their own
each. Without the variable they are skipped; CI starts PostgreSQL 16 for
them, and a test fails the run when the variable is missing there.

Operations. ops/README.md describes the database: the Unix socket with peer
authentication, so that no password exists, the one-time preparation, and
rollback across a schema step. The service unit starts after PostgreSQL.
docs/LIMITATIONS.md no longer lists the in-memory store and now lists the
missing backup and the missing deletion of accounts.
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
The account API kept everything in the memory of its process, so every
restart and every deployment of the API deleted all accounts and signed
everyone out.

PostgresAccountStore implements IAccountStore on PostgreSQL through Npgsql.
The three uniqueness rules are unique constraints, so two requests that race
for one name, address or Google identity are decided by the database. A
refused account is an ordinary answer, not an exception, and the rules are
reported in the order the in-memory store reports them. A session refers to
its account, and a session of an account that is not stored is refused; the
in-memory store now refuses it as well.

The schema is a list of numbered steps. The service takes the missing ones
when it starts, in one transaction under an advisory lock, and does not
start when the database cannot be reached or is newer than the release.
/health answers 503 while the database does not answer.

ConnectionStrings:Accounts selects PostgreSQL. A production service that is
given no database does not start; keeping accounts in memory there has to
be asked for with Accounts:Store=Memory. Development and the tests keep the
in-memory store by default.

The store gains RemoveExpiredSessionsAsync, and a background service calls
it at the start and once an hour: a session whose token is never presented
again no longer keeps its row for ever.

PostgreSQL keeps a timestamp to the microsecond. The store cuts a moment to
that precision and returns what it stored, and the service uses the
returned account, so the creation time in the registration response equals
the one every later read returns.

Tests. AccountStoreContractTests states the rules once and runs them on
both stores. Further tests cover the schema steps applied concurrently and a
database that is too new, the choice of the store from the configuration,
the removal of expired sessions, and an account with its session surviving
a restart of the service over HTTP. The PostgreSQL tests read their
database from PROGMAWEB_TEST_POSTGRES and work in a schema of their own
each. Without the variable they are skipped; CI starts PostgreSQL 16 for
them, and a test fails the run when the variable is missing there.

Operations. ops/README.md describes the database: the Unix socket with peer
authentication, so that no password exists, the one-time preparation, and
rollback across a schema step. The service unit starts after PostgreSQL.
docs/LIMITATIONS.md no longer lists the in-memory store and now lists the
missing backup and the missing deletion of accounts.
</pre>
</div>
</content>
</entry>
</feed>
